{"about":{"site":"https://codewithpapers.app","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.","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"},"url":"/code/run-eval","entry":"run_eval","source":"Syntology graph, per-sample; not an archive number","read_at":"2026-09-24T18:15:14+00:00","claim":"Names are grouped by exact entry-name string. Same-named routines are NOT asserted to be equivalent; 'ran' means executed on a synthesized fixture, not correctness. n_samples_ran = sum of by_status over every status except 'unverified' (ran_draft_wrong and ran_fixture are failures of Syntology's instrument, not of the code); n_papers_ran = papers with at least one such sample.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"},"n_papers":12,"n_papers_ran":4,"units":"n_samples, n_samples_ran, n_samples_fingerprinted and by_status count distinct code bodies (code_sha256); n_places and n_places_pointer_only count places, one per (paper, code body) pair, which is also the unit of the samples list","n_samples":10,"n_samples_ran":3,"n_samples_fingerprinted":0,"n_places":12,"n_places_pointer_only":2,"by_status":{"ran_honours":0,"ran_violates":0,"ran_draft_wrong":1,"ran_fixture":0,"ran":2,"unverified":7},"syntology":{"atlas_url":null,"mcp":null,"mcp_per_sample":{"tool":"get_code","arguments_in":"samples[].mcp_get_code"},"developers":"https://syntology.ai/developers"},"samples":[{"arxiv_id":"2609.02991","paper":"/paper/arxiv-2609-02991","title":"TRACE: Spatiotemporal Contact Memory Graph Network Simulator for Granular Dynamics","date":null,"month_inferred_from_arxiv_id":"2026-09","title_source":"syntology","repo":"Data-Driven-Computational-Geotechnics/TRACE","path":"common/eval_common.py","file_url":"https://github.com/Data-Driven-Computational-Geotechnics/TRACE/blob/HEAD/common/eval_common.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"155a7ca64cbef7ca","mcp_get_code":{"code_sha256":"155a7ca64cbef7ca"}},{"arxiv_id":"2608.00656","paper":"/paper/arxiv-2608-00656","title":"Simulation-Based Plate-Reverb Parameter Estimation from a Single Impulse Response","date":null,"month_inferred_from_arxiv_id":"2026-08","title_source":"syntology","repo":"LOGUNIVPM/1st-DAFx-Challenge","path":"batchEval.py","file_url":"https://github.com/LOGUNIVPM/1st-DAFx-Challenge/blob/HEAD/batchEval.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"4ddf9028d1984a85","mcp_get_code":{"code_sha256":"4ddf9028d1984a85"}},{"arxiv_id":"2605.20075","paper":"/paper/arxiv-2605-20075","title":"CopT: Contrastive On-Policy Thinking with Continuous Spaces for General and Agentic Reasoning","date":null,"month_inferred_from_arxiv_id":"2026-05","title_source":"syntology","repo":"shishirpatil/gorilla","path":"gorilla/inference/gorilla_eval.py","file_url":"https://github.com/shishirpatil/gorilla/blob/HEAD/gorilla/inference/gorilla_eval.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"0739c6a004099445","mcp_get_code":{"code_sha256":"0739c6a004099445"}},{"arxiv_id":"2510.19176","paper":"/paper/arxiv-2510-19176","title":"The Zero-Step Thinking: An Empirical Study of Mode Selection as Harder Early Exit in Reasoning Models","date":null,"month_inferred_from_arxiv_id":"2025-10","title_source":"syntology","repo":"Trae1ounG/Zero_Step_Thinking","path":"src/train_predictor_with_class_weights.py","file_url":"https://github.com/Trae1ounG/Zero_Step_Thinking/blob/HEAD/src/train_predictor_with_class_weights.py","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"56958f10c524ec4e","mcp_get_code":{"code_sha256":"56958f10c524ec4e"}},{"arxiv_id":"2404.06921","paper":"/paper/goex-perspectives-and-designs-towards-a","title":"GoEX: Perspectives and Designs Towards a Runtime for Autonomous LLM Applications","date":"2024-04-10","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"ShishirPatil/gorilla","path":"gorilla/inference/gorilla_eval.py","file_url":"https://github.com/ShishirPatil/gorilla/blob/HEAD/gorilla/inference/gorilla_eval.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"0739c6a004099445","mcp_get_code":{"code_sha256":"0739c6a004099445"}},{"arxiv_id":"2401.10727","paper":"/paper/tool-lmm-a-large-multi-modal-model-for-tool","title":"MLLM-Tool: A Multimodal Large Language Model For Tool Agent Learning","date":"2024-01-19","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"mllm-tool/mllm-tool","path":"code/inference.py","file_url":"https://github.com/mllm-tool/mllm-tool/blob/HEAD/code/inference.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"22a14acf7a13a6b0","mcp_get_code":{"code_sha256":"22a14acf7a13a6b0"}},{"arxiv_id":"2401.00563","paper":"/paper/kernelgpt-enhanced-kernel-fuzzing-via-large","title":"KernelGPT: Enhanced Kernel Fuzzing via Large Language Models","date":"2023-12-31","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"ise-uiuc/kernelgpt","path":"spec-gen/eval_spec.py","file_url":"https://github.com/ise-uiuc/kernelgpt/blob/HEAD/spec-gen/eval_spec.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"4b04b35a66f40278","mcp_get_code":{"code_sha256":"4b04b35a66f40278"}},{"arxiv_id":"2310.16111","paper":"/paper/locally-differentially-private-document","title":"Locally Differentially Private Document Generation Using Zero Shot Prompting","date":"2023-10-24","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"saitejautpala/dp_prompt","path":"attacks/text_attacker_architecture.py","file_url":"https://github.com/saitejautpala/dp_prompt/blob/HEAD/attacks/text_attacker_architecture.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"cb16209dce94db5f","mcp_get_code":{"code_sha256":"cb16209dce94db5f"}},{"arxiv_id":"2201.01549","paper":"/paper/spt-code-sequence-to-sequence-pre-training","title":"SPT-Code: Sequence-to-Sequence Pre-Training for Learning Source Code Representations","date":null,"month_inferred_from_arxiv_id":"2022-01","title_source":"archive","repo":"nougatca/spt-code","path":"sources/downstream_tasks/search.py","file_url":"https://github.com/nougatca/spt-code/blob/HEAD/sources/downstream_tasks/search.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"1646803daed79ebe","mcp_get_code":{"code_sha256":"1646803daed79ebe"}},{"arxiv_id":"2009.07364","paper":"/paper/an-information-theoretic-view-on-selecting","title":"An information theoretic view on selecting linguistic probes","date":"2020-09-15","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"SPOClab-ca/InfoProbe","path":"src/pos_probe.py","file_url":"https://github.com/SPOClab-ca/InfoProbe/blob/HEAD/src/pos_probe.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"f8435c21eef9039f","mcp_get_code":{"code_sha256":"f8435c21eef9039f"}},{"arxiv_id":"2024.findings-acl.39","paper":null,"title":"arXiv:2024.findings-acl.39","date":null,"month_inferred_from_arxiv_id":null,"title_source":null,"repo":"passing2961/DribeR","path":"open_vlm_main.py","file_url":"https://github.com/passing2961/DribeR/blob/HEAD/open_vlm_main.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"970a21f1c0fb22ef","mcp_get_code":{"code_sha256":"970a21f1c0fb22ef"}},{"arxiv_id":"2023.findings-emnlp.566","paper":null,"title":"arXiv:2023.findings-emnlp.566","date":null,"month_inferred_from_arxiv_id":null,"title_source":null,"repo":"SaitejaUtpala/dp_prompt","path":"attacks/text_attacker_architecture.py","file_url":"https://github.com/SaitejaUtpala/dp_prompt/blob/HEAD/attacks/text_attacker_architecture.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"cb16209dce94db5f","mcp_get_code":{"code_sha256":"cb16209dce94db5f"}}]}