{"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/sort-key","entry":"sort_key","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":9,"n_papers_ran":5,"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":8,"n_samples_ran":5,"n_samples_fingerprinted":0,"n_places":9,"n_places_pointer_only":5,"by_status":{"ran_honours":0,"ran_violates":0,"ran_draft_wrong":1,"ran_fixture":0,"ran":4,"unverified":3},"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":"2605.31584","paper":"/paper/arxiv-2605-31584","title":"LongTraceRL: Learning Long-Context Reasoning from Search Agent Trajectories with Rubric Rewards","date":null,"month_inferred_from_arxiv_id":"2026-05","title_source":"syntology","repo":"THU-KEG/LongTraceRL","path":"slime/slime/ray/placement_group.py","file_url":"https://github.com/THU-KEG/LongTraceRL/blob/HEAD/slime/slime/ray/placement_group.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"f0bef468d87d7f58","mcp_get_code":{"code_sha256":"f0bef468d87d7f58"}},{"arxiv_id":"2604.00986","paper":"/paper/arxiv-2604-00986","title":"Do Phone-Use Agents Respect Your Privacy?","date":null,"month_inferred_from_arxiv_id":"2026-04","title_source":"syntology","repo":"FreedomIntelligence/MyPhoneBench","path":"android_world/checkpointer.py","file_url":"https://github.com/FreedomIntelligence/MyPhoneBench/blob/HEAD/android_world/checkpointer.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":"fb545d6f19945d37","mcp_get_code":{"code_sha256":"fb545d6f19945d37"}},{"arxiv_id":"2507.01299","paper":null,"title":"arXiv:2507.01299","date":null,"month_inferred_from_arxiv_id":"2025-07","title_source":null,"repo":"alibaba/EfficientAI","path":"dcore/paper/cal_bfcl_lazy_reasoning.py","file_url":"https://github.com/alibaba/EfficientAI/blob/HEAD/dcore/paper/cal_bfcl_lazy_reasoning.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":"7ef1fef589b5e4a2","mcp_get_code":{"code_sha256":"7ef1fef589b5e4a2"}},{"arxiv_id":"2409.15867","paper":"/paper/in-context-ensemble-improves-video-language","title":"In-Context Ensemble Learning from Pseudo Labels Improves Video-Language Models for Low-Level Workflow Understanding","date":"2024-09-24","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"moucheng2017/action-labelling","path":"helpers.py","file_url":"https://github.com/moucheng2017/action-labelling/blob/HEAD/helpers.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"ee3a0742a3013971","mcp_get_code":{"code_sha256":"ee3a0742a3013971"}},{"arxiv_id":"2405.14573","paper":"/paper/androidworld-a-dynamic-benchmarking","title":"AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents","date":"2024-05-23","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"google-research/android_world","path":"android_world/checkpointer.py","file_url":"https://github.com/google-research/android_world/blob/HEAD/android_world/checkpointer.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":false,"code_sha256_prefix":"fb545d6f19945d37","mcp_get_code":{"code_sha256":"fb545d6f19945d37"}},{"arxiv_id":"2401.13246","paper":"/paper/seer-facilitating-structured-reasoning-and","title":"SEER: Facilitating Structured Reasoning and Explanation via Reinforcement Learning","date":"2024-01-24","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"Chen-GX/SEER","path":"etree_task1/src/RL_etree.py","file_url":"https://github.com/Chen-GX/SEER/blob/HEAD/etree_task1/src/RL_etree.py","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"5218be161f3c9262","mcp_get_code":{"code_sha256":"5218be161f3c9262"}},{"arxiv_id":"2312.10302","paper":"/paper/one-shot-learning-as-instruction-data","title":"One-Shot Learning as Instruction Data Prospector for Large Language Models","date":"2023-12-16","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"pldlgb/nuggets","path":"sort.py","file_url":"https://github.com/pldlgb/nuggets/blob/HEAD/sort.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"ffeeafe3f7000afc","mcp_get_code":{"code_sha256":"ffeeafe3f7000afc"}},{"arxiv_id":"2112.04888","paper":"/paper/a-bilingual-openworld-video-text-dataset-and","title":"A Bilingual, OpenWorld Video Text Dataset and End-to-end Video Text Spotter with Transformer","date":"2021-12-09","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"weijiawu/transvtspotter","path":"models/save_track.py","file_url":"https://github.com/weijiawu/transvtspotter/blob/HEAD/models/save_track.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"eedc05e12aec5bd3","mcp_get_code":{"code_sha256":"eedc05e12aec5bd3"}},{"arxiv_id":"1612.01744","paper":"/paper/listen-and-translate-a-proof-of-concept-for","title":"Listen and Translate: A Proof of Concept for End-to-End Speech-to-Text Translation","date":"2016-12-06","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"eske/seq2seq","path":"config/AMU/avg.py","file_url":"https://github.com/eske/seq2seq/blob/HEAD/config/AMU/avg.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":"94dc1f1512737035","mcp_get_code":{"code_sha256":"94dc1f1512737035"}}]}