{"about":{"non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","site":"https://codewithpapers.app","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/method/multi-head-attention/papers/ran/6","list_of":"/method/multi-head-attention","method":"Multi-Head Attention","archive":{"snapshot":"2025-07-28"},"syntology_read_at":"2026-09-28T10:30:06+00:00","order":"ran","order_definition":"only papers where Syntology ran at least one harvested sample; date (newest first), ties by arXiv id","caption":"We ran code from the paper's repository; we did not isolate this method inside it.","absence":"A paper missing from this list is not a recorded non-run: it may have no arXiv id, no harvested code, or only samples that have not run yet.","page":6,"pages_in_order":35,"rows_per_page":100,"rows":[501,600],"of":3454,"counts":{"archive_papers_tagged":24855,"with_a_code_link":11214,"where_syntology_ran_a_sample":3454,"not_listed_spam_title":0,"listed":24855,"listed_where_code_ran":3454,"where_syntology_ran_a_sample_split":{"with_a_run_with_no_instrument_failure":2916,"every_run_a_failure_of_syntologys_instrument":538,"listed_with_a_run_with_no_instrument_failure":2916,"listed_every_run_a_failure_of_syntologys_instrument":538,"filter":{"states":["a run with no instrument failure","any run, instrument failures included"],"default":"a run with no instrument failure","note":"on the 'only where code ran' pages the default hides, in the browser, the rows where every run was a failure of Syntology's instrument; the second state shows them again. Rows are hidden, never re-ordered; these twins list every row"}},"definition":"distinct papers the archive tags; 'where Syntology ran a sample' counts papers with at least one harvested sample that ran, which is not a correctness claim"},"first_page":"/method/multi-head-attention/papers/ran/1","prev":"/method/multi-head-attention/papers/ran/5","next":"/method/multi-head-attention/papers/ran/7","papers":[{"paper":"/paper/sop-unlock-the-power-of-social-facilitation","slug":"sop-unlock-the-power-of-social-facilitation","title":"SeqAR: Jailbreak LLMs with Sequential Auto-Generated Characters","date":"2024-07-02","arxiv_id":"2407.01902","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":1,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["yang-yan-yang-yan/sop"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/multilingual-trolley-problems-for-language","slug":"multilingual-trolley-problems-for-language","title":"Language Model Alignment in Multilingual Trolley Problems","date":"2024-07-02","arxiv_id":"2407.02273","n_code_links":2,"syntology":{"ran":3,"of":4,"n_ran_checked":3,"n_instrument":0,"unverified":1,"pointer_only":0,"phrase":"3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["causalNLP/moralmachine","causalnlp/multitp"],"state":"official (archive's flag): 3 ran","n_ran":3,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/the-art-of-saying-no-contextual-noncompliance","slug":"the-art-of-saying-no-contextual-noncompliance","title":"The Art of Saying No: Contextual Noncompliance in Language Models","date":"2024-07-02","arxiv_id":"2407.12043","n_code_links":1,"syntology":{"ran":4,"of":5,"n_ran_checked":4,"n_instrument":0,"unverified":1,"pointer_only":1,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":null}},{"paper":"/paper/bergen-a-benchmarking-library-for-retrieval","slug":"bergen-a-benchmarking-library-for-retrieval","title":"BERGEN: A Benchmarking Library for Retrieval-Augmented Generation","date":"2024-07-01","arxiv_id":"2407.01102","n_code_links":1,"syntology":{"ran":2,"of":2,"n_ran_checked":0,"n_instrument":2,"unverified":0,"pointer_only":2,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified","official":{"repos":["naver/bergen"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/searching-for-best-practices-in-retrieval","slug":"searching-for-best-practices-in-retrieval","title":"Searching for Best Practices in Retrieval-Augmented Generation","date":"2024-07-01","arxiv_id":"2407.01219","n_code_links":1,"syntology":{"ran":2,"of":2,"n_ran_checked":2,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["FudanDNN-NLP/RAG"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":2,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/hypformer-exploring-efficient-hyperbolic","slug":"hypformer-exploring-efficient-hyperbolic","title":"Hypformer: Exploring Efficient Hyperbolic Transformer Fully in Hyperbolic Space","date":"2024-07-01","arxiv_id":"2407.01290","n_code_links":1,"syntology":{"ran":5,"of":7,"n_ran_checked":5,"n_instrument":0,"unverified":2,"pointer_only":1,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified","official":{"repos":["Graph-and-Geometric-Learning/hyperbolic-transformer"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":5,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":"/paper/increasing-model-capacity-for-free-a-simple","slug":"increasing-model-capacity-for-free-a-simple","title":"Increasing Model Capacity for Free: A Simple Strategy for Parameter Efficient Fine-tuning","date":"2024-07-01","arxiv_id":"2407.01320","n_code_links":1,"syntology":{"ran":7,"of":10,"n_ran_checked":6,"n_instrument":1,"unverified":3,"pointer_only":7,"phrase":"7 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 1 where Syntology's instrument failed) · 3 unverified","official":{"repos":["lins-lab/capaboost"],"state":"official (archive's flag): 7 ran","n_ran":7,"n_constructed":0,"n_ran_no_instrument_failure":6,"n_unverified":3,"ran_from_kinds":["official"]}}},{"paper":"/paper/summary-of-a-haystack-a-challenge-to-long","slug":"summary-of-a-haystack-a-challenge-to-long","title":"Summary of a Haystack: A Challenge to Long-Context LLMs and RAG Systems","date":"2024-07-01","arxiv_id":"2407.01370","n_code_links":1,"syntology":{"ran":2,"of":2,"n_ran_checked":0,"n_instrument":2,"unverified":0,"pointer_only":0,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified","official":{"repos":["salesforce/summary-of-a-haystack"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/retrieval-augmented-generation-in","slug":"retrieval-augmented-generation-in","title":"Retrieval-augmented generation in multilingual settings","date":"2024-07-01","arxiv_id":"2407.01463","n_code_links":1,"syntology":{"ran":2,"of":2,"n_ran_checked":0,"n_instrument":2,"unverified":0,"pointer_only":2,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified","official":{"repos":["naver/bergen"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/deciphering-the-factors-influencing-the","slug":"deciphering-the-factors-influencing-the","title":"Deciphering the Factors Influencing the Efficacy of Chain-of-Thought: Probability, Memorization, and Noisy Reasoning","date":"2024-07-01","arxiv_id":"2407.01687","n_code_links":1,"syntology":{"ran":1,"of":4,"n_ran_checked":1,"n_instrument":0,"unverified":3,"pointer_only":4,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified","official":{"repos":["aksh555/deciphering_cot"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":3,"ran_from_kinds":["official"]}}},{"paper":"/paper/towards-universal-mesh-movement-networks","slug":"towards-universal-mesh-movement-networks","title":"Towards Universal Mesh Movement Networks","date":"2024-06-29","arxiv_id":"2407.00382","n_code_links":1,"syntology":{"ran":3,"of":7,"n_ran_checked":1,"n_instrument":2,"unverified":4,"pointer_only":0,"phrase":"3 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 2 where Syntology's instrument failed) · 4 unverified","official":{"repos":["mesh-adaptation/um2n"],"state":"official (archive's flag): 3 ran","n_ran":3,"n_constructed":1,"n_ran_no_instrument_failure":1,"n_unverified":4,"ran_from_kinds":["official"]}}},{"paper":"/paper/infinigen-efficient-generative-inference-of","slug":"infinigen-efficient-generative-inference-of","title":"InfiniGen: Efficient Generative Inference of Large Language Models with Dynamic KV Cache Management","date":"2024-06-28","arxiv_id":"2406.19707","n_code_links":1,"syntology":{"ran":14,"of":15,"n_ran_checked":10,"n_instrument":4,"unverified":1,"pointer_only":1,"phrase":"14 ran (of which 0 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 0 violated, 10 with no contract checked; 4 where Syntology's instrument failed) · 1 unverified","official":null}},{"paper":"/paper/shortcutsbench-a-large-scale-real-world","slug":"shortcutsbench-a-large-scale-real-world","title":"ShortcutsBench: A Large-Scale Real-world Benchmark for API-based Agents","date":"2024-06-28","arxiv_id":"2407.00132","n_code_links":1,"syntology":{"ran":6,"of":7,"n_ran_checked":4,"n_instrument":2,"unverified":1,"pointer_only":0,"phrase":"6 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 1 violated, 3 with no contract checked; 2 where Syntology's instrument failed) · 1 unverified","official":{"repos":["eachsheep/shortcutsbench"],"state":"official (archive's flag): 6 ran","n_ran":6,"n_constructed":0,"n_ran_no_instrument_failure":4,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/multimodal-prototyping-for-cancer-survival","slug":"multimodal-prototyping-for-cancer-survival","title":"Multimodal Prototyping for cancer survival prediction","date":"2024-06-28","arxiv_id":"2407.00224","n_code_links":1,"syntology":{"ran":7,"of":12,"n_ran_checked":4,"n_instrument":3,"unverified":5,"pointer_only":12,"phrase":"7 ran (of which 3 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 3 where Syntology's instrument failed) · 5 unverified","official":{"repos":["mahmoodlab/MMP"],"state":"official (archive's flag): 7 ran","n_ran":7,"n_constructed":3,"n_ran_no_instrument_failure":4,"n_unverified":5,"ran_from_kinds":["official"]}}},{"paper":"/paper/sonnet-or-not-bot-poetry-evaluation-for-large","slug":"sonnet-or-not-bot-poetry-evaluation-for-large","title":"Sonnet or Not, Bot? Poetry Evaluation for Large Models and Datasets","date":"2024-06-27","arxiv_id":"2406.18906","n_code_links":1,"syntology":{"ran":3,"of":3,"n_ran_checked":2,"n_instrument":1,"unverified":0,"pointer_only":3,"phrase":"3 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 2 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","official":{"repos":["maria-antoniak/poetry-eval"],"state":"official (archive's flag): 3 ran","n_ran":3,"n_constructed":0,"n_ran_no_instrument_failure":2,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/unigen-a-unified-framework-for-textual","slug":"unigen-a-unified-framework-for-textual","title":"UniGen: A Unified Framework for Textual Dataset Generation Using Large Language Models","date":"2024-06-27","arxiv_id":"2406.18966","n_code_links":1,"syntology":{"ran":8,"of":13,"n_ran_checked":8,"n_instrument":0,"unverified":5,"pointer_only":13,"phrase":"8 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 0 where Syntology's instrument failed) · 5 unverified","official":{"repos":["howiehwong/unigen"],"state":"official (archive's flag): 8 ran","n_ran":8,"n_constructed":0,"n_ran_no_instrument_failure":8,"n_unverified":5,"ran_from_kinds":["official"]}}},{"paper":"/paper/seakr-self-aware-knowledge-retrieval-for","slug":"seakr-self-aware-knowledge-retrieval-for","title":"SeaKR: Self-aware Knowledge Retrieval for Adaptive Retrieval Augmented Generation","date":"2024-06-27","arxiv_id":"2406.19215","n_code_links":1,"syntology":{"ran":10,"of":12,"n_ran_checked":9,"n_instrument":1,"unverified":2,"pointer_only":12,"phrase":"10 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 1 where Syntology's instrument failed) · 2 unverified","official":{"repos":["thu-keg/seakr"],"state":"official (archive's flag): 10 ran","n_ran":10,"n_constructed":0,"n_ran_no_instrument_failure":9,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":"/paper/human-aware-vision-and-language-navigation","slug":"human-aware-vision-and-language-navigation","title":"Human-Aware Vision-and-Language Navigation: Bridging Simulation to Reality with Dynamic Human Interactions","date":"2024-06-27","arxiv_id":"2406.19236","n_code_links":1,"syntology":{"ran":9,"of":10,"n_ran_checked":9,"n_instrument":0,"unverified":1,"pointer_only":10,"phrase":"9 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["lpercc/ha3d_simulator"],"state":"official (archive's flag): 7 ran","n_ran":7,"n_constructed":0,"n_ran_no_instrument_failure":7,"n_unverified":1,"ran_from_kinds":["found_in_text","official"]}}},{"paper":"/paper/mfdnet-multi-frequency-deflare-network-for","slug":"mfdnet-multi-frequency-deflare-network-for","title":"MFDNet: Multi-Frequency Deflare Network for Efficient Nighttime Flare Removal","date":"2024-06-26","arxiv_id":"2406.18079","n_code_links":1,"syntology":{"ran":6,"of":7,"n_ran_checked":4,"n_instrument":2,"unverified":1,"pointer_only":7,"phrase":"6 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 2 where Syntology's instrument failed) · 1 unverified","official":{"repos":["Jiang-maomao/flare-removal"],"state":"official (archive's flag): 6 ran","n_ran":6,"n_constructed":0,"n_ran_no_instrument_failure":4,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/mathodyssey-benchmarking-mathematical-problem","slug":"mathodyssey-benchmarking-mathematical-problem","title":"MathOdyssey: Benchmarking Mathematical Problem-Solving Skills in Large Language Models Using Odyssey Math Data","date":"2024-06-26","arxiv_id":"2406.18321","n_code_links":3,"syntology":{"ran":5,"of":5,"n_ran_checked":4,"n_instrument":1,"unverified":0,"pointer_only":0,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","official":null}},{"paper":"/paper/themis-towards-flexible-and-interpretable-nlg","slug":"themis-towards-flexible-and-interpretable-nlg","title":"Themis: A Reference-free NLG Evaluation Language Model with Flexibility and Interpretability","date":"2024-06-26","arxiv_id":"2406.18365","n_code_links":1,"syntology":{"ran":5,"of":5,"n_ran_checked":3,"n_instrument":2,"unverified":0,"pointer_only":0,"phrase":"5 ran (of which 3 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified","official":{"repos":["PKU-ONELab/Themis"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":3,"n_ran_no_instrument_failure":3,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/wildguard-open-one-stop-moderation-tools-for","slug":"wildguard-open-one-stop-moderation-tools-for","title":"WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs","date":"2024-06-26","arxiv_id":"2406.18495","n_code_links":4,"syntology":{"ran":1,"of":3,"n_ran_checked":1,"n_instrument":0,"unverified":2,"pointer_only":3,"phrase":"1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified; the one sample that ran constructed an object rather than computing a result","official":{"repos":["allenai/wildguard"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":1,"n_ran_no_instrument_failure":1,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":"/paper/step-dpo-step-wise-preference-optimization","slug":"step-dpo-step-wise-preference-optimization","title":"Step-DPO: Step-wise Preference Optimization for Long-chain Reasoning of LLMs","date":"2024-06-26","arxiv_id":"2406.18629","n_code_links":1,"syntology":{"ran":11,"of":12,"n_ran_checked":10,"n_instrument":1,"unverified":1,"pointer_only":12,"phrase":"11 ran (of which 0 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 0 violated, 10 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified","official":{"repos":["dvlab-research/step-dpo"],"state":"official (archive's flag): 11 ran","n_ran":11,"n_constructed":0,"n_ran_no_instrument_failure":10,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/understand-what-llm-needs-dual-preference","slug":"understand-what-llm-needs-dual-preference","title":"Understand What LLM Needs: Dual Preference Alignment for Retrieval-Augmented Generation","date":"2024-06-26","arxiv_id":"2406.18676","n_code_links":1,"syntology":{"ran":1,"of":3,"n_ran_checked":1,"n_instrument":0,"unverified":2,"pointer_only":3,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified","official":{"repos":["dongguanting/dpa-rag"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":"/paper/jailbreaking-llms-with-arabic-transliteration","slug":"jailbreaking-llms-with-arabic-transliteration","title":"Jailbreaking LLMs with Arabic Transliteration and Arabizi","date":"2024-06-26","arxiv_id":"2406.18725","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["securedl/arabic_jailbreak"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/a-stem-agnostic-single-decoder-system-for","slug":"a-stem-agnostic-single-decoder-system-for","title":"A Stem-Agnostic Single-Decoder System for Music Source Separation Beyond Four Stems","date":"2024-06-26","arxiv_id":"2406.18747","n_code_links":1,"syntology":{"ran":4,"of":4,"n_ran_checked":4,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["kwatcharasupat/query-bandit"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":4,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/unlocking-continual-learning-abilities-in","slug":"unlocking-continual-learning-abilities-in","title":"Unlocking Continual Learning Abilities in Language Models","date":"2024-06-25","arxiv_id":"2406.17245","n_code_links":1,"syntology":{"ran":5,"of":7,"n_ran_checked":5,"n_instrument":0,"unverified":2,"pointer_only":0,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified","official":{"repos":["wenyudu/migu"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":5,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":"/paper/q-dit-accurate-post-training-quantization-for","slug":"q-dit-accurate-post-training-quantization-for","title":"Q-DiT: Accurate Post-Training Quantization for Diffusion Transformers","date":"2024-06-25","arxiv_id":"2406.17343","n_code_links":1,"syntology":{"ran":17,"of":19,"n_ran_checked":11,"n_instrument":6,"unverified":2,"pointer_only":19,"phrase":"17 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 4 honoured, 0 violated, 7 with no contract checked; 6 where Syntology's instrument failed) · 2 unverified","official":{"repos":["juanerx/q-dit"],"state":"official (archive's flag): 16 ran","n_ran":16,"n_constructed":0,"n_ran_no_instrument_failure":11,"n_unverified":2,"ran_from_kinds":["official","unlocated"]}}},{"paper":"/paper/lumberchunker-long-form-narrative-document","slug":"lumberchunker-long-form-narrative-document","title":"LumberChunker: Long-Form Narrative Document Segmentation","date":"2024-06-25","arxiv_id":"2406.17526","n_code_links":1,"syntology":{"ran":2,"of":2,"n_ran_checked":1,"n_instrument":1,"unverified":0,"pointer_only":2,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","official":{"repos":["joaodsmarques/lumberchunker"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/director3d-real-world-camera-trajectory-and","slug":"director3d-real-world-camera-trajectory-and","title":"Director3D: Real-world Camera Trajectory and 3D Scene Generation from Text","date":"2024-06-25","arxiv_id":"2406.17601","n_code_links":1,"syntology":{"ran":12,"of":15,"n_ran_checked":7,"n_instrument":5,"unverified":3,"pointer_only":15,"phrase":"12 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 3 honoured, 0 violated, 4 with no contract checked; 5 where Syntology's instrument failed) · 3 unverified","official":{"repos":["imlixinyang/director3d"],"state":"official (archive's flag): 12 ran","n_ran":12,"n_constructed":0,"n_ran_no_instrument_failure":7,"n_unverified":3,"ran_from_kinds":["official"]}}},{"paper":"/paper/structured-unrestricted-rank-matrices-for","slug":"structured-unrestricted-rank-matrices-for","title":"Structured Unrestricted-Rank Matrices for Parameter Efficient Fine-tuning","date":"2024-06-25","arxiv_id":"2406.17740","n_code_links":1,"syntology":{"ran":2,"of":2,"n_ran_checked":1,"n_instrument":1,"unverified":0,"pointer_only":0,"phrase":"2 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","official":{"repos":["arijitthegame/structured-matrices-peft"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":1,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/sum-saliency-unification-through-mamba-for","slug":"sum-saliency-unification-through-mamba-for","title":"SUM: Saliency Unification through Mamba for Visual Attention Modeling","date":"2024-06-25","arxiv_id":"2406.17815","n_code_links":1,"syntology":{"ran":14,"of":14,"n_ran_checked":14,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"14 ran (of which 0 constructed an object rather than computing a result; 14 with no instrument failure: 0 honoured, 0 violated, 14 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["Arhosseini77/SUM"],"state":"official (archive's flag): 14 ran","n_ran":14,"n_constructed":0,"n_ran_no_instrument_failure":14,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/finding-transformer-circuits-with-edge","slug":"finding-transformer-circuits-with-edge","title":"Finding Transformer Circuits with Edge Pruning","date":"2024-06-24","arxiv_id":"2406.16778","n_code_links":1,"syntology":{"ran":8,"of":8,"n_ran_checked":3,"n_instrument":5,"unverified":0,"pointer_only":0,"phrase":"8 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 3 honoured, 0 violated, 0 with no contract checked; 5 where Syntology's instrument failed) · 0 unverified","official":{"repos":["princeton-nlp/edge-pruning"],"state":"official (archive's flag): 8 ran","n_ran":8,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/ragnarok-a-reusable-rag-framework-and","slug":"ragnarok-a-reusable-rag-framework-and","title":"Ragnarök: A Reusable RAG Framework and Baselines for TREC 2024 Retrieval-Augmented Generation Track","date":"2024-06-24","arxiv_id":"2406.16828","n_code_links":2,"syntology":{"ran":19,"of":23,"n_ran_checked":19,"n_instrument":0,"unverified":4,"pointer_only":0,"phrase":"19 ran (of which 0 constructed an object rather than computing a result; 19 with no instrument failure: 0 honoured, 0 violated, 19 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified","official":{"repos":["castorini/ragnarok"],"state":"official (archive's flag): 9 ran","n_ran":9,"n_constructed":0,"n_ran_no_instrument_failure":9,"n_unverified":3,"ran_from_kinds":["listed","official"]}}},{"paper":"/paper/dreambench-a-human-aligned-benchmark-for","slug":"dreambench-a-human-aligned-benchmark-for","title":"DreamBench++: A Human-Aligned Benchmark for Personalized Image Generation","date":"2024-06-24","arxiv_id":"2406.16855","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["yuangpeng/dreambench_plus"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/breaking-the-frame-image-retrieval-by-visual","slug":"breaking-the-frame-image-retrieval-by-visual","title":"Breaking the Frame: Visual Place Recognition by Overlap Prediction","date":"2024-06-23","arxiv_id":"2406.16204","n_code_links":1,"syntology":{"ran":4,"of":5,"n_ran_checked":4,"n_instrument":0,"unverified":1,"pointer_only":5,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["weitong8591/vop"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":4,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/ladder-a-model-agnostic-framework-boosting","slug":"ladder-a-model-agnostic-framework-boosting","title":"Ladder: A Model-Agnostic Framework Boosting LLM-based Machine Translation to the Next Level","date":"2024-06-22","arxiv_id":"2406.15741","n_code_links":3,"syntology":{"ran":2,"of":8,"n_ran_checked":0,"n_instrument":2,"unverified":6,"pointer_only":8,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 6 unverified","official":{"repos":["fzp0424/ladder","fzp0424/mt-ladder"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":6,"ran_from_kinds":["official"]}}},{"paper":"/paper/fast-tree-field-integrators-from-low","slug":"fast-tree-field-integrators-from-low","title":"Fast Tree-Field Integrators: From Low Displacement Rank to Topological Transformers","date":"2024-06-22","arxiv_id":"2406.15881","n_code_links":1,"syntology":{"ran":5,"of":5,"n_ran_checked":0,"n_instrument":5,"unverified":0,"pointer_only":5,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 5 where Syntology's instrument failed) · 0 unverified","official":{"repos":["brcsomnath/fasttreeintegrator"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/augmenting-query-and-passage-for-retrieval","slug":"augmenting-query-and-passage-for-retrieval","title":"QPaug: Question and Passage Augmentation for Open-Domain Question Answering of LLMs","date":"2024-06-20","arxiv_id":"2406.14277","n_code_links":1,"syntology":{"ran":2,"of":3,"n_ran_checked":2,"n_instrument":0,"unverified":1,"pointer_only":3,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["kmswin1/qpaug"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":2,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/learning-to-plan-for-retrieval-augmented","slug":"learning-to-plan-for-retrieval-augmented","title":"Learning to Plan for Retrieval-Augmented Large Language Models from Knowledge Graphs","date":"2024-06-20","arxiv_id":"2406.14282","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":1,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["zjukg/lpkg"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/coderag-bench-can-retrieval-augment-code","slug":"coderag-bench-can-retrieval-augment-code","title":"CodeRAG-Bench: Can Retrieval Augment Code Generation?","date":"2024-06-20","arxiv_id":"2406.14497","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":1,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["code-rag-bench/code-rag-bench"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/llasa-large-multimodal-agent-for-human","slug":"llasa-large-multimodal-agent-for-human","title":"LLaSA: A Multimodal LLM for Human Activity Analysis Through Wearable and Smartphone Sensors","date":"2024-06-20","arxiv_id":"2406.14498","n_code_links":1,"syntology":{"ran":6,"of":7,"n_ran_checked":3,"n_instrument":3,"unverified":1,"pointer_only":7,"phrase":"6 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 1 violated, 2 with no contract checked; 3 where Syntology's instrument failed) · 1 unverified","official":{"repos":["bashlab/llasa"],"state":"official (archive's flag): 6 ran","n_ran":6,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/mmbench-video-a-long-form-multi-shot","slug":"mmbench-video-a-long-form-multi-shot","title":"MMBench-Video: A Long-Form Multi-Shot Benchmark for Holistic Video Understanding","date":"2024-06-20","arxiv_id":"2406.14515","n_code_links":1,"syntology":{"ran":5,"of":6,"n_ran_checked":1,"n_instrument":4,"unverified":1,"pointer_only":0,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 1 unverified","official":{"repos":["open-compass/vlmevalkit"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/prism-a-framework-for-decoupling-and","slug":"prism-a-framework-for-decoupling-and","title":"Prism: A Framework for Decoupling and Assessing the Capabilities of VLMs","date":"2024-06-20","arxiv_id":"2406.14544","n_code_links":1,"syntology":{"ran":9,"of":13,"n_ran_checked":9,"n_instrument":0,"unverified":4,"pointer_only":0,"phrase":"9 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified","official":{"repos":["sparksjoe/prism"],"state":"official (archive's flag): 9 ran","n_ran":9,"n_constructed":0,"n_ran_no_instrument_failure":9,"n_unverified":4,"ran_from_kinds":["official"]}}},{"paper":"/paper/how-to-compute-the-probability-of-a-word","slug":"how-to-compute-the-probability-of-a-word","title":"How to Compute the Probability of a Word","date":"2024-06-20","arxiv_id":"2406.14561","n_code_links":2,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["tpimentelms/probability-of-a-word"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/evaluating-rag-fusion-with-ragelo-an","slug":"evaluating-rag-fusion-with-ragelo-an","title":"Evaluating RAG-Fusion with RAGElo: an Automated Elo-based Framework","date":"2024-06-20","arxiv_id":"2406.14783","n_code_links":1,"syntology":{"ran":3,"of":4,"n_ran_checked":3,"n_instrument":0,"unverified":1,"pointer_only":0,"phrase":"3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["zetaalphavector/ragelo"],"state":"official (archive's flag): 3 ran","n_ran":3,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/can-long-context-language-models-subsume","slug":"can-long-context-language-models-subsume","title":"Can Long-Context Language Models Subsume Retrieval, RAG, SQL, and More?","date":"2024-06-19","arxiv_id":"2406.13121","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["google-deepmind/loft"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/patholm-identifying-pathogenicity-from-the","slug":"patholm-identifying-pathogenicity-from-the","title":"PathoLM: Identifying pathogenicity from the DNA sequence through the Genome Foundation Model","date":"2024-06-19","arxiv_id":"2406.13133","n_code_links":1,"syntology":{"ran":2,"of":2,"n_ran_checked":0,"n_instrument":2,"unverified":0,"pointer_only":0,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified","official":{"repos":["Sajib-006/Patho-LM"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/efficient-sharpness-aware-minimization-for-2","slug":"efficient-sharpness-aware-minimization-for-2","title":"Efficient Sharpness-Aware Minimization for Molecular Graph Transformer Models","date":"2024-06-19","arxiv_id":"2406.13137","n_code_links":1,"syntology":{"ran":7,"of":7,"n_ran_checked":7,"n_instrument":0,"unverified":0,"pointer_only":7,"phrase":"7 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["YL-wang/GraphSAM"],"state":"official (archive's flag): 7 ran","n_ran":7,"n_constructed":0,"n_ran_no_instrument_failure":7,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/r-2ag-incorporating-retrieval-information","slug":"r-2ag-incorporating-retrieval-information","title":"R^2AG: Incorporating Retrieval Information into Retrieval Augmented Generation","date":"2024-06-19","arxiv_id":"2406.13249","n_code_links":1,"syntology":{"ran":2,"of":4,"n_ran_checked":0,"n_instrument":2,"unverified":2,"pointer_only":0,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified","official":{"repos":["yefd/RRAG"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":"/paper/do-multimodal-foundation-models-understand","slug":"do-multimodal-foundation-models-understand","title":"WONDERBREAD: A Benchmark for Evaluating Multimodal Foundation Models on Business Process Management Tasks","date":"2024-06-19","arxiv_id":"2406.13264","n_code_links":1,"syntology":{"ran":11,"of":12,"n_ran_checked":11,"n_instrument":0,"unverified":1,"pointer_only":12,"phrase":"11 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 0 violated, 11 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["hazyresearch/wonderbread"],"state":"official (archive's flag): 11 ran","n_ran":11,"n_constructed":0,"n_ran_no_instrument_failure":11,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/part-aware-unified-representation-of-language-1","slug":"part-aware-unified-representation-of-language-1","title":"Part-aware Unified Representation of Language and Skeleton for Zero-shot Action Recognition","date":"2024-06-19","arxiv_id":"2406.13327","n_code_links":1,"syntology":{"ran":3,"of":3,"n_ran_checked":3,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 1 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["azzh1/purls"],"state":"official (archive's flag): 3 ran","n_ran":3,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/instructrag-instructing-retrieval-augmented","slug":"instructrag-instructing-retrieval-augmented","title":"InstructRAG: Instructing Retrieval-Augmented Generation via Self-Synthesized Rationales","date":"2024-06-19","arxiv_id":"2406.13629","n_code_links":1,"syntology":{"ran":2,"of":3,"n_ran_checked":0,"n_instrument":2,"unverified":1,"pointer_only":0,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 1 unverified","official":{"repos":["weizhepei/instructrag"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/model-internals-based-answer-attribution-for","slug":"model-internals-based-answer-attribution-for","title":"Model Internals-based Answer Attribution for Trustworthy Retrieval-Augmented Generation","date":"2024-06-19","arxiv_id":"2406.13663","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["betswish/mirage"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/alanavlm-a-multimodal-embodied-ai-foundation","slug":"alanavlm-a-multimodal-embodied-ai-foundation","title":"AlanaVLM: A Multimodal Embodied AI Foundation Model for Egocentric Video Understanding","date":"2024-06-19","arxiv_id":"2406.13807","n_code_links":1,"syntology":{"ran":4,"of":4,"n_ran_checked":4,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["alanaai/evud"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":4,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/hierarchical-associative-memory-parallelized","slug":"hierarchical-associative-memory-parallelized","title":"Hierarchical Associative Memory, Parallelized MLP-Mixer, and Symmetry Breaking","date":"2024-06-18","arxiv_id":"2406.12220","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["Toshihiro-Ota/paramixer"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/towards-a-client-centered-assessment-of-llm","slug":"towards-a-client-centered-assessment-of-llm","title":"Towards a Client-Centered Assessment of LLM Therapists by Client Simulation","date":"2024-06-18","arxiv_id":"2406.12266","n_code_links":1,"syntology":{"ran":2,"of":2,"n_ran_checked":0,"n_instrument":2,"unverified":0,"pointer_only":2,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified","official":{"repos":["wangjs9/clientcast"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/an-investigation-of-neuron-activation-as-a","slug":"an-investigation-of-neuron-activation-as-a","title":"An Investigation of Neuron Activation as a Unified Lens to Explain Chain-of-Thought Eliciting Arithmetic Reasoning of LLMs","date":"2024-06-18","arxiv_id":"2406.12288","n_code_links":2,"syntology":{"ran":34,"of":36,"n_ran_checked":32,"n_instrument":2,"unverified":2,"pointer_only":12,"phrase":"34 ran (of which 0 constructed an object rather than computing a result; 32 with no instrument failure: 0 honoured, 3 violated, 29 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified","official":{"repos":["dakingrai/neuron-analysis-cot-arithmetic-reasoning"],"state":"official (archive's flag): 11 ran","n_ran":11,"n_constructed":0,"n_ran_no_instrument_failure":9,"n_unverified":0,"ran_from_kinds":["listed","official"]}}},{"paper":"/paper/planrag-a-plan-then-retrieval-augmented","slug":"planrag-a-plan-then-retrieval-augmented","title":"PlanRAG: A Plan-then-Retrieval Augmented Generation for Generative Large Language Models as Decision Makers","date":"2024-06-18","arxiv_id":"2406.12430","n_code_links":1,"syntology":{"ran":5,"of":6,"n_ran_checked":5,"n_instrument":0,"unverified":1,"pointer_only":0,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["myeon9h/planrag"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":5,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/judging-the-judges-evaluating-alignment-and","slug":"judging-the-judges-evaluating-alignment-and","title":"Judging the Judges: Evaluating Alignment and Vulnerabilities in LLMs-as-Judges","date":"2024-06-18","arxiv_id":"2406.12624","n_code_links":1,"syntology":{"ran":2,"of":3,"n_ran_checked":2,"n_instrument":0,"unverified":1,"pointer_only":3,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["UMass-Meta-LLM-Eval/llm_eval"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":2,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/chatglm-a-family-of-large-language-models","slug":"chatglm-a-family-of-large-language-models","title":"ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools","date":"2024-06-18","arxiv_id":"2406.12793","n_code_links":7,"syntology":{"ran":21,"of":29,"n_ran_checked":20,"n_instrument":1,"unverified":8,"pointer_only":1,"phrase":"21 ran (of which 0 constructed an object rather than computing a result; 20 with no instrument failure: 0 honoured, 0 violated, 20 with no contract checked; 1 where Syntology's instrument failed) · 8 unverified","official":{"repos":["thudm/chatglm-6b"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":1,"ran_from_kinds":["listed","official"]}}},{"paper":"/paper/can-large-language-models-always-solve-easy","slug":"can-large-language-models-always-solve-easy","title":"Can Large Language Models Always Solve Easy Problems if They Can Solve Harder Ones?","date":"2024-06-18","arxiv_id":"2406.12809","n_code_links":1,"syntology":{"ran":11,"of":14,"n_ran_checked":11,"n_instrument":0,"unverified":3,"pointer_only":4,"phrase":"11 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 0 violated, 11 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified","official":{"repos":["QwenLM/ConsisEval"],"state":"official (archive's flag): 11 ran","n_ran":11,"n_constructed":0,"n_ran_no_instrument_failure":11,"n_unverified":3,"ran_from_kinds":["official"]}}},{"paper":"/paper/interpretable-preferences-via-multi-objective","slug":"interpretable-preferences-via-multi-objective","title":"Interpretable Preferences via Multi-Objective Reward Modeling and Mixture-of-Experts","date":"2024-06-18","arxiv_id":"2406.12845","n_code_links":2,"syntology":{"ran":9,"of":9,"n_ran_checked":9,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"9 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["RLHFlow/RLHF-Reward-Modeling"],"state":"community repositories only","n_ran":0,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["listed"]}}},{"paper":"/paper/dart-math-difficulty-aware-rejection-tuning-1","slug":"dart-math-difficulty-aware-rejection-tuning-1","title":"DART-Math: Difficulty-Aware Rejection Tuning for Mathematical Problem-Solving","date":"2024-06-18","arxiv_id":"2407.13690","n_code_links":1,"syntology":{"ran":10,"of":12,"n_ran_checked":10,"n_instrument":0,"unverified":2,"pointer_only":0,"phrase":"10 ran (of which 0 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 0 violated, 10 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified","official":{"repos":["hkust-nlp/dart-math"],"state":"official (archive's flag): 10 ran","n_ran":10,"n_constructed":0,"n_ran_no_instrument_failure":10,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":"/paper/textit-refiner-restructure-retrieval-content","slug":"textit-refiner-restructure-retrieval-content","title":"Refiner: Restructure Retrieval Content Efficiently to Advance Question-Answering Capabilities","date":"2024-06-17","arxiv_id":"2406.11357","n_code_links":1,"syntology":{"ran":7,"of":8,"n_ran_checked":6,"n_instrument":1,"unverified":1,"pointer_only":8,"phrase":"7 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified","official":{"repos":["allen-li1231/refiner-rag"],"state":"official (archive's flag): 7 ran","n_ran":7,"n_constructed":0,"n_ran_no_instrument_failure":6,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/trace-the-evidence-constructing-knowledge","slug":"trace-the-evidence-constructing-knowledge","title":"TRACE the Evidence: Constructing Knowledge-Grounded Reasoning Chains for Retrieval-Augmented Generation","date":"2024-06-17","arxiv_id":"2406.11460","n_code_links":2,"syntology":{"ran":4,"of":4,"n_ran_checked":0,"n_instrument":4,"unverified":0,"pointer_only":4,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 0 unverified","official":{"repos":["jyfang6/trace"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/geogpt4v-towards-geometric-multi-modal-large","slug":"geogpt4v-towards-geometric-multi-modal-large","title":"GeoGPT4V: Towards Geometric Multi-modal Large Language Models with Geometric Image Generation","date":"2024-06-17","arxiv_id":"2406.11503","n_code_links":1,"syntology":{"ran":2,"of":2,"n_ran_checked":0,"n_instrument":2,"unverified":0,"pointer_only":0,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified","official":{"repos":["lanyu0303/geogpt4v_project"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/scaling-the-codebook-size-of-vqgan-to-100000","slug":"scaling-the-codebook-size-of-vqgan-to-100000","title":"Scaling the Codebook Size of VQGAN to 100,000 with a Utilization Rate of 99%","date":"2024-06-17","arxiv_id":"2406.11837","n_code_links":1,"syntology":{"ran":2,"of":2,"n_ran_checked":1,"n_instrument":1,"unverified":0,"pointer_only":2,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","official":{"repos":["zh460045050/vqgan-lc"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official","unlocated"]}}},{"paper":"/paper/fintruthqa-a-benchmark-dataset-for-evaluating","slug":"fintruthqa-a-benchmark-dataset-for-evaluating","title":"FinTruthQA: A Benchmark Dataset for Evaluating the Quality of Financial Information Disclosure","date":"2024-06-17","arxiv_id":"2406.12009","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["bethxx99/FinTruthQA"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/citrus-chunked-instruction-aware-state","slug":"citrus-chunked-instruction-aware-state","title":"CItruS: Chunked Instruction-aware State Eviction for Long Sequence Modeling","date":"2024-06-17","arxiv_id":"2406.12018","n_code_links":1,"syntology":{"ran":5,"of":10,"n_ran_checked":0,"n_instrument":5,"unverified":5,"pointer_only":10,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 5 where Syntology's instrument failed) · 5 unverified","official":{"repos":["ybai-nlp/CItruS"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":5,"ran_from_kinds":["official"]}}},{"paper":"/paper/welldunn-on-the-robustness-and-explainability","slug":"welldunn-on-the-robustness-and-explainability","title":"WellDunn: On the Robustness and Explainability of Language Models and Large Language Models in Identifying Wellness Dimensions","date":"2024-06-17","arxiv_id":"2406.12058","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":1,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["vedantpalit/WellDunn"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/sharelora-parameter-efficient-and-robust","slug":"sharelora-parameter-efficient-and-robust","title":"ShareLoRA: Parameter Efficient and Robust Large Language Model Fine-tuning via Shared Low-Rank Adaptation","date":"2024-06-16","arxiv_id":"2406.10785","n_code_links":1,"syntology":{"ran":5,"of":6,"n_ran_checked":0,"n_instrument":5,"unverified":1,"pointer_only":6,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified","official":{"repos":["Rain9876/ShareLoRA"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/blockpruner-fine-grained-pruning-for-large","slug":"blockpruner-fine-grained-pruning-for-large","title":"BlockPruner: Fine-grained Pruning for Large Language Models","date":"2024-06-15","arxiv_id":"2406.10594","n_code_links":1,"syntology":{"ran":9,"of":12,"n_ran_checked":6,"n_instrument":3,"unverified":3,"pointer_only":12,"phrase":"9 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 1 violated, 5 with no contract checked; 3 where Syntology's instrument failed) · 3 unverified","official":{"repos":["MrGGLS/BlockPruner"],"state":"official (archive's flag): 9 ran","n_ran":9,"n_constructed":0,"n_ran_no_instrument_failure":6,"n_unverified":3,"ran_from_kinds":["official"]}}},{"paper":"/paper/when-will-gradient-regularization-be-harmful","slug":"when-will-gradient-regularization-be-harmful","title":"When Will Gradient Regularization Be Harmful?","date":"2024-06-14","arxiv_id":"2406.09723","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":0,"n_instrument":1,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","official":{"repos":["zhaoyang-0204/gnp"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/blend-a-benchmark-for-llms-on-everyday","slug":"blend-a-benchmark-for-llms-on-everyday","title":"BLEnD: A Benchmark for LLMs on Everyday Knowledge in Diverse Cultures and Languages","date":"2024-06-14","arxiv_id":"2406.09948","n_code_links":1,"syntology":{"ran":1,"of":10,"n_ran_checked":0,"n_instrument":1,"unverified":9,"pointer_only":10,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 9 unverified","official":{"repos":["nlee0212/blend"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":9,"ran_from_kinds":["official"]}}},{"paper":"/paper/neural-concept-binder","slug":"neural-concept-binder","title":"Neural Concept Binder","date":"2024-06-14","arxiv_id":"2406.09949","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["ml-research/neuralconceptbinder"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/know-the-unknown-an-uncertainty-sensitive","slug":"know-the-unknown-an-uncertainty-sensitive","title":"Know the Unknown: An Uncertainty-Sensitive Method for LLM Instruction Tuning","date":"2024-06-14","arxiv_id":"2406.10099","n_code_links":1,"syntology":{"ran":4,"of":6,"n_ran_checked":4,"n_instrument":0,"unverified":2,"pointer_only":0,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified","official":{"repos":["jiaqili404/trustworthyrag"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":4,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":"/paper/liere-generalizing-rotary-position-encodings","slug":"liere-generalizing-rotary-position-encodings","title":"LieRE: Generalizing Rotary Position Encodings","date":"2024-06-14","arxiv_id":"2406.10322","n_code_links":1,"syntology":{"ran":1,"of":3,"n_ran_checked":1,"n_instrument":0,"unverified":2,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified","official":{"repos":["stanford-aimi/liere"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":"/paper/denoisereid-denoising-model-for","slug":"denoisereid-denoising-model-for","title":"DenoiseRep: Denoising Model for Representation Learning","date":"2024-06-13","arxiv_id":"2406.08773","n_code_links":1,"syntology":{"ran":17,"of":22,"n_ran_checked":17,"n_instrument":0,"unverified":5,"pointer_only":6,"phrase":"17 ran (of which 2 constructed an object rather than computing a result; 17 with no instrument failure: 2 honoured, 1 violated, 14 with no contract checked; 0 where Syntology's instrument failed) · 5 unverified","official":{"repos":["wangguanan/denoiserep"],"state":"official (archive's flag): 17 ran","n_ran":17,"n_constructed":2,"n_ran_no_instrument_failure":17,"n_unverified":5,"ran_from_kinds":["official"]}}},{"paper":"/paper/fredformer-frequency-debiased-transformer-for","slug":"fredformer-frequency-debiased-transformer-for","title":"Fredformer: Frequency Debiased Transformer for Time Series Forecasting","date":"2024-06-13","arxiv_id":"2406.09009","n_code_links":1,"syntology":{"ran":1,"of":2,"n_ran_checked":1,"n_instrument":0,"unverified":1,"pointer_only":2,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["chenzrg/fredformer"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/cross-modal-learning-for-anomaly-detection-in","slug":"cross-modal-learning-for-anomaly-detection-in","title":"Cross-Modal Learning for Anomaly Detection in Complex Industrial Process: Methodology and Benchmark","date":"2024-06-13","arxiv_id":"2406.09016","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["gaochangwu/fmf-benchmark"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/jailbreakeval-an-integrated-toolkit-for","slug":"jailbreakeval-an-integrated-toolkit-for","title":"JailbreakEval: An Integrated Toolkit for Evaluating Jailbreak Attempts Against Large Language Models","date":"2024-06-13","arxiv_id":"2406.09321","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["thuccslab/jailbreakeval"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/alleviating-distortion-in-image-generation","slug":"alleviating-distortion-in-image-generation","title":"Alleviating Distortion in Image Generation via Multi-Resolution Diffusion Models and Time-Dependent Layer Normalization","date":"2024-06-13","arxiv_id":"2406.09416","n_code_links":1,"syntology":{"ran":13,"of":16,"n_ran_checked":11,"n_instrument":2,"unverified":3,"pointer_only":6,"phrase":"13 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 0 violated, 11 with no contract checked; 2 where Syntology's instrument failed) · 3 unverified","official":{"repos":["qihao067/DiMR"],"state":"official (archive's flag): 13 ran","n_ran":13,"n_constructed":0,"n_ran_no_instrument_failure":11,"n_unverified":3,"ran_from_kinds":["official"]}}},{"paper":"/paper/judging-the-judges-a-systematic-investigation","slug":"judging-the-judges-a-systematic-investigation","title":"Judging the Judges: A Systematic Study of Position Bias in LLM-as-a-Judge","date":"2024-06-12","arxiv_id":"2406.07791","n_code_links":1,"syntology":{"ran":16,"of":16,"n_ran_checked":16,"n_instrument":0,"unverified":0,"pointer_only":16,"phrase":"16 ran (of which 0 constructed an object rather than computing a result; 16 with no instrument failure: 0 honoured, 0 violated, 16 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["Slimshilin/Position-Bias-Analyzer-Demo"],"state":"official (archive's flag): 16 ran","n_ran":16,"n_constructed":0,"n_ran_no_instrument_failure":16,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/mail-improving-imitation-learning-with-mamba","slug":"mail-improving-imitation-learning-with-mamba","title":"MaIL: Improving Imitation Learning with Mamba","date":"2024-06-12","arxiv_id":"2406.08234","n_code_links":1,"syntology":{"ran":1,"of":2,"n_ran_checked":1,"n_instrument":0,"unverified":1,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["alrhub/mail"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/fine-tuned-small-llms-still-significantly","slug":"fine-tuned-small-llms-still-significantly","title":"Fine-Tuned 'Small' LLMs (Still) Significantly Outperform Zero-Shot Generative AI Models in Text Classification","date":"2024-06-12","arxiv_id":"2406.08660","n_code_links":1,"syntology":{"ran":5,"of":9,"n_ran_checked":5,"n_instrument":0,"unverified":4,"pointer_only":9,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified","official":{"repos":["mnbucher/text-cls-llms"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":5,"n_unverified":4,"ran_from_kinds":["official"]}}},{"paper":"/paper/dara-decomposition-alignment-reasoning","slug":"dara-decomposition-alignment-reasoning","title":"DARA: Decomposition-Alignment-Reasoning Autonomous Language Agent for Question Answering over Knowledge Graphs","date":"2024-06-11","arxiv_id":"2406.07080","n_code_links":1,"syntology":{"ran":9,"of":9,"n_ran_checked":9,"n_instrument":0,"unverified":0,"pointer_only":9,"phrase":"9 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["UKPLab/acl2024-DARA"],"state":"official (archive's flag): 9 ran","n_ran":9,"n_constructed":0,"n_ran_no_instrument_failure":9,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/ai-sandbagging-language-models-can","slug":"ai-sandbagging-language-models-can","title":"AI Sandbagging: Language Models can Strategically Underperform on Evaluations","date":"2024-06-11","arxiv_id":"2406.07358","n_code_links":1,"syntology":{"ran":5,"of":5,"n_ran_checked":5,"n_instrument":0,"unverified":0,"pointer_only":5,"phrase":"5 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["teunvdweij/sandbagging"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":5,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/mllmguard-a-multi-dimensional-safety","slug":"mllmguard-a-multi-dimensional-safety","title":"MLLMGuard: A Multi-dimensional Safety Evaluation Suite for Multimodal Large Language Models","date":"2024-06-11","arxiv_id":"2406.07594","n_code_links":1,"syntology":{"ran":4,"of":4,"n_ran_checked":4,"n_instrument":0,"unverified":0,"pointer_only":4,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["Carol-gutianle/MLLMGuard"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":4,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/data-efficient-learning-with-neural-programs","slug":"data-efficient-learning-with-neural-programs","title":"Data-Efficient Learning with Neural Programs","date":"2024-06-10","arxiv_id":"2406.06246","n_code_links":1,"syntology":{"ran":4,"of":6,"n_ran_checked":4,"n_instrument":0,"unverified":2,"pointer_only":0,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified","official":{"repos":["alaiasolkobreslin/ised"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":4,"n_unverified":2,"ran_from_kinds":["official"]}}},{"paper":"/paper/compute-better-spent-replacing-dense-layers","slug":"compute-better-spent-replacing-dense-layers","title":"Compute Better Spent: Replacing Dense Layers with Structured Matrices","date":"2024-06-10","arxiv_id":"2406.06248","n_code_links":1,"syntology":{"ran":4,"of":5,"n_ran_checked":4,"n_instrument":0,"unverified":1,"pointer_only":0,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["shikaiqiu/compute-better-spent"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":4,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/husky-a-unified-open-source-language-agent","slug":"husky-a-unified-open-source-language-agent","title":"Husky: A Unified, Open-Source Language Agent for Multi-Step Reasoning","date":"2024-06-10","arxiv_id":"2406.06469","n_code_links":1,"syntology":{"ran":6,"of":6,"n_ran_checked":3,"n_instrument":3,"unverified":0,"pointer_only":6,"phrase":"6 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 1 honoured, 2 violated, 0 with no contract checked; 3 where Syntology's instrument failed) · 0 unverified","official":{"repos":["agent-husky/husky-v1"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":0,"ran_from_kinds":["official","unlocated"]}}},{"paper":"/paper/domainrag-a-chinese-benchmark-for-evaluating","slug":"domainrag-a-chinese-benchmark-for-evaluating","title":"DomainRAG: A Chinese Benchmark for Evaluating Domain-specific Retrieval-Augmented Generation","date":"2024-06-09","arxiv_id":"2406.05654","n_code_links":2,"syntology":{"ran":12,"of":12,"n_ran_checked":11,"n_instrument":1,"unverified":0,"pointer_only":12,"phrase":"12 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 0 violated, 11 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","official":{"repos":["ShootingWong/DomainRAG"],"state":"official (archive's flag): 12 ran","n_ran":12,"n_constructed":0,"n_ran_no_instrument_failure":11,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/sinklora-enhanced-efficiency-and-chat","slug":"sinklora-enhanced-efficiency-and-chat","title":"SinkLoRA: Enhanced Efficiency and Chat Capabilities for Long-Context Large Language Models","date":"2024-06-09","arxiv_id":"2406.05678","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":0,"n_instrument":1,"unverified":0,"pointer_only":1,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","official":{"repos":["dexter-gt-86/sinklora"],"state":"official: no sample here; runs from other or unrecorded repositories","n_ran":0,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["unlocated"]}}},{"paper":"/paper/attention-as-a-hypernetwork","slug":"attention-as-a-hypernetwork","title":"Attention as a Hypernetwork","date":"2024-06-09","arxiv_id":"2406.05816","n_code_links":1,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","official":{"repos":["smonsays/hypernetwork-attention"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/transformer-conformal-prediction-for-time","slug":"transformer-conformal-prediction-for-time","title":"Transformer Conformal Prediction for Time Series","date":"2024-06-08","arxiv_id":"2406.05332","n_code_links":1,"syntology":{"ran":3,"of":3,"n_ran_checked":0,"n_instrument":3,"unverified":0,"pointer_only":3,"phrase":"3 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 3 where Syntology's instrument failed) · 0 unverified","official":{"repos":["Jayaos/TCPTS"],"state":"official (archive's flag): 3 ran","n_ran":3,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/a-fine-tuning-dataset-and-benchmark-for-large","slug":"a-fine-tuning-dataset-and-benchmark-for-large","title":"A Fine-tuning Dataset and Benchmark for Large Language Models for Protein Understanding","date":"2024-06-08","arxiv_id":"2406.05540","n_code_links":1,"syntology":{"ran":6,"of":7,"n_ran_checked":6,"n_instrument":0,"unverified":1,"pointer_only":0,"phrase":"6 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["tsynbio/proteinlmdataset"],"state":"official (archive's flag): 6 ran","n_ran":6,"n_constructed":0,"n_ran_no_instrument_failure":6,"n_unverified":1,"ran_from_kinds":["official"]}}},{"paper":"/paper/mixture-of-agents-enhances-large-language","slug":"mixture-of-agents-enhances-large-language","title":"Mixture-of-Agents Enhances Large Language Model Capabilities","date":"2024-06-07","arxiv_id":"2406.04692","n_code_links":3,"syntology":{"ran":3,"of":5,"n_ran_checked":1,"n_instrument":2,"unverified":2,"pointer_only":3,"phrase":"3 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified","official":null}},{"paper":"/paper/crag-comprehensive-rag-benchmark","slug":"crag-comprehensive-rag-benchmark","title":"CRAG -- Comprehensive RAG Benchmark","date":"2024-06-07","arxiv_id":"2406.04744","n_code_links":2,"syntology":{"ran":4,"of":4,"n_ran_checked":3,"n_instrument":1,"unverified":0,"pointer_only":4,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","official":{"repos":["facebookresearch/crag"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/berts-are-generative-in-context-learners","slug":"berts-are-generative-in-context-learners","title":"BERTs are Generative In-Context Learners","date":"2024-06-07","arxiv_id":"2406.04823","n_code_links":1,"syntology":{"ran":13,"of":26,"n_ran_checked":12,"n_instrument":1,"unverified":13,"pointer_only":0,"phrase":"13 ran (of which 0 constructed an object rather than computing a result; 12 with no instrument failure: 0 honoured, 1 violated, 11 with no contract checked; 1 where Syntology's instrument failed) · 13 unverified","official":{"repos":["ltgoslo/bert-in-context"],"state":"official (archive's flag): 13 ran","n_ran":13,"n_constructed":0,"n_ran_no_instrument_failure":12,"n_unverified":13,"ran_from_kinds":["official"]}}}],"record_sha256":"d3fcff49bbbbf1b34fa000cc5f7980a7117f902b18163cf91c5a12baba4fd692","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}