{"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/process-response","entry":"process_response","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":18,"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":13,"n_samples_ran":5,"n_samples_fingerprinted":2,"n_places":23,"n_places_pointer_only":8,"by_status":{"ran_honours":0,"ran_violates":0,"ran_draft_wrong":2,"ran_fixture":0,"ran":3,"unverified":8},"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":"2605.31058","paper":"/paper/arxiv-2605-31058","title":"Combinatorial Synthesis: Scaling Code RLVR via Atomic Decomposition and Recombination","date":null,"month_inferred_from_arxiv_id":"2026-05","title_source":"syntology","repo":"icip-cas/ADR","path":"0_pipeline.py","file_url":"https://github.com/icip-cas/ADR/blob/HEAD/0_pipeline.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":"455e1ebd4a1d8af1","mcp_get_code":{"code_sha256":"455e1ebd4a1d8af1"}},{"arxiv_id":"2603.21461","paper":"/paper/arxiv-2603-21461","title":"DSPA: Dynamic SAE Steering for Data-Efficient Preference Alignment","date":null,"month_inferred_from_arxiv_id":"2026-03","title_source":"syntology","repo":"lm-sys/FastChat","path":"fastchat/model/model_chatglm.py","file_url":"https://github.com/lm-sys/FastChat/blob/HEAD/fastchat/model/model_chatglm.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":"dc21db7cf25bfe72","mcp_get_code":{"code_sha256":"dc21db7cf25bfe72"}},{"arxiv_id":"2503.17352","paper":"/paper/openvlthinker-an-early-exploration-to-complex","title":"OpenVLThinker: An Early Exploration to Complex Vision-Language Reasoning via Iterative Self-Improvement","date":"2025-03-21","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"yihedeng9/openvlthinker","path":"v1/evaluation/eval_openvlthinker.py","file_url":"https://github.com/yihedeng9/openvlthinker/blob/HEAD/v1/evaluation/eval_openvlthinker.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":"394985bb469413a6","mcp_get_code":{"code_sha256":"394985bb469413a6"}},{"arxiv_id":"2503.17352","paper":"/paper/openvlthinker-an-early-exploration-to-complex","title":"OpenVLThinker: An Early Exploration to Complex Vision-Language Reasoning via Iterative Self-Improvement","date":"2025-03-21","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"yihedeng9/openvlthinker","path":"v1/evaluation/eval_qwen.py","file_url":"https://github.com/yihedeng9/openvlthinker/blob/HEAD/v1/evaluation/eval_qwen.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":"10457e62fda12718","mcp_get_code":{"code_sha256":"10457e62fda12718"}},{"arxiv_id":"2501.11120","paper":"/paper/tell-me-about-yourself-llms-are-aware-of","title":"Tell me about yourself: LLMs are aware of their learned behaviors","date":"2025-01-19","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"xuchanbao/behavioral-self-awareness","path":"code/multiple-choice/evaluation/self-report-free-apples.py","file_url":"https://github.com/xuchanbao/behavioral-self-awareness/blob/HEAD/code/multiple-choice/evaluation/self-report-free-apples.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"9ae68598a33c4eca","mcp_get_code":{"code_sha256":"9ae68598a33c4eca"}},{"arxiv_id":"2501.11120","paper":"/paper/tell-me-about-yourself-llms-are-aware-of","title":"Tell me about yourself: LLMs are aware of their learned behaviors","date":"2025-01-19","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"xuchanbao/behavioral-self-awareness","path":"code/multiple-choice/evaluation/self-report-free-for-final-plots.py","file_url":"https://github.com/xuchanbao/behavioral-self-awareness/blob/HEAD/code/multiple-choice/evaluation/self-report-free-for-final-plots.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"5e3a9bb9f06658ca","mcp_get_code":{"code_sha256":"5e3a9bb9f06658ca"}},{"arxiv_id":"2501.11120","paper":"/paper/tell-me-about-yourself-llms-are-aware-of","title":"Tell me about yourself: LLMs are aware of their learned behaviors","date":"2025-01-19","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"xuchanbao/behavioral-self-awareness","path":"code/multiple-choice/evaluation/self-report-free-myopia.py","file_url":"https://github.com/xuchanbao/behavioral-self-awareness/blob/HEAD/code/multiple-choice/evaluation/self-report-free-myopia.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"34a401b2e4a8163a","mcp_get_code":{"code_sha256":"34a401b2e4a8163a"}},{"arxiv_id":"2408.16500","paper":"/paper/cogvlm2-visual-language-models-for-image-and","title":"CogVLM2: Visual Language Models for Image and Video Understanding","date":"2024-08-29","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"thudm/glm-4","path":"demo/composite_demo/src/client.py","file_url":"https://github.com/thudm/glm-4/blob/HEAD/demo/composite_demo/src/client.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"51ffd8516205ff45","mcp_get_code":{"code_sha256":"51ffd8516205ff45"}},{"arxiv_id":"2406.14670","paper":"/paper/exploring-design-choices-for-building","title":"Exploring Design Choices for Building Language-Specific LLMs","date":"2024-06-20","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"atutej/token-language-adaptation","path":"FastChat/fastchat/model/model_chatglm.py","file_url":"https://github.com/atutej/token-language-adaptation/blob/HEAD/FastChat/fastchat/model/model_chatglm.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"dc21db7cf25bfe72","mcp_get_code":{"code_sha256":"dc21db7cf25bfe72"}},{"arxiv_id":"2406.12793","paper":"/paper/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","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"thudm/chatglm4","path":"demo/composite_demo/src/client.py","file_url":"https://github.com/thudm/chatglm4/blob/HEAD/demo/composite_demo/src/client.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"51ffd8516205ff45","mcp_get_code":{"code_sha256":"51ffd8516205ff45"}},{"arxiv_id":"2404.00699","paper":"/paper/how-much-are-llms-contaminated-a","title":"How Much are Large Language Models Contaminated? A Comprehensive Survey and the LLMSanitize Library","date":"2024-03-31","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"ntunlp/llmsanitize","path":"llmsanitize/closed_data_methods/ts_guessing_question_based.py","file_url":"https://github.com/ntunlp/llmsanitize/blob/HEAD/llmsanitize/closed_data_methods/ts_guessing_question_based.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":"be50c8c31c5cd309","mcp_get_code":{"code_sha256":"be50c8c31c5cd309"}},{"arxiv_id":"2404.00699","paper":"/paper/how-much-are-llms-contaminated-a","title":"How Much are Large Language Models Contaminated? A Comprehensive Survey and the LLMSanitize Library","date":"2024-03-31","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"ntunlp/llmsanitize","path":"llmsanitize/closed_data_methods/ts_guessing_question_multichoice.py","file_url":"https://github.com/ntunlp/llmsanitize/blob/HEAD/llmsanitize/closed_data_methods/ts_guessing_question_multichoice.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":"4abd4eaf3320d068","mcp_get_code":{"code_sha256":"4abd4eaf3320d068"}},{"arxiv_id":"2403.04132","paper":"/paper/chatbot-arena-an-open-platform-for-evaluating","title":"Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference","date":"2024-03-07","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"BirgerMoell/SwedishLLMBenchmark","path":"fastchat/model/model_chatglm.py","file_url":"https://github.com/BirgerMoell/SwedishLLMBenchmark/blob/HEAD/fastchat/model/model_chatglm.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":"dc21db7cf25bfe72","mcp_get_code":{"code_sha256":"dc21db7cf25bfe72"}},{"arxiv_id":"2403.02502","paper":"/paper/trial-and-error-exploration-based-trajectory","title":"Trial and Error: Exploration-Based Trajectory Optimization for LLM Agents","date":"2024-03-04","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"yifan-song793/eto","path":"fastchat/model/model_chatglm.py","file_url":"https://github.com/yifan-song793/eto/blob/HEAD/fastchat/model/model_chatglm.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"dc21db7cf25bfe72","mcp_get_code":{"code_sha256":"dc21db7cf25bfe72"}},{"arxiv_id":"2402.06627","paper":"/paper/feedback-loops-with-language-models-drive-in","title":"Feedback Loops With Language Models Drive In-Context Reward Hacking","date":"2024-02-09","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"aypan17/llm-feedback","path":"output-refinement/optimization/utils.py","file_url":"https://github.com/aypan17/llm-feedback/blob/HEAD/output-refinement/optimization/utils.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"676f96b3b5ebf7e4","mcp_get_code":{"code_sha256":"676f96b3b5ebf7e4"}},{"arxiv_id":"2402.06627","paper":"/paper/feedback-loops-with-language-models-drive-in","title":"Feedback Loops With Language Models Drive In-Context Reward Hacking","date":"2024-02-09","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"aypan17/llm-feedback","path":"output-refinement/reward_hacking/utils.py","file_url":"https://github.com/aypan17/llm-feedback/blob/HEAD/output-refinement/reward_hacking/utils.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"5747fd8a748faaad","mcp_get_code":{"code_sha256":"5747fd8a748faaad"}},{"arxiv_id":"2402.02130","paper":"/paper/rendering-graphs-for-graph-reasoning-in","title":"GITA: Graph to Visual and Textual Integration for Vision-Language Graph Reasoning","date":"2024-02-03","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"WEIYanbin1999/GITA","path":"fastchat/model/model_chatglm.py","file_url":"https://github.com/WEIYanbin1999/GITA/blob/HEAD/fastchat/model/model_chatglm.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"dc21db7cf25bfe72","mcp_get_code":{"code_sha256":"dc21db7cf25bfe72"}},{"arxiv_id":"2402.01830","paper":"/paper/peer-review-in-llms-automatic-evaluation","title":"PiCO: Peer Review in LLMs based on the Consistency Optimization","date":"2024-02-02","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"PKU-YuanGroup/Peer-review-in-LLMs","path":"llm_judge/config/model_chatglm.py","file_url":"https://github.com/PKU-YuanGroup/Peer-review-in-LLMs/blob/HEAD/llm_judge/config/model_chatglm.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"dc21db7cf25bfe72","mcp_get_code":{"code_sha256":"dc21db7cf25bfe72"}},{"arxiv_id":"2310.17631","paper":"/paper/judgelm-fine-tuned-large-language-models-are","title":"JudgeLM: Fine-tuned Large Language Models are Scalable Judges","date":"2023-10-26","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"hitz-zentroa/eval-MCG-COLING-2025","path":"evaluation/JudgeLM-main/judgelm/model/chatglm_model.py","file_url":"https://github.com/hitz-zentroa/eval-MCG-COLING-2025/blob/HEAD/evaluation/JudgeLM-main/judgelm/model/chatglm_model.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":"dc21db7cf25bfe72","mcp_get_code":{"code_sha256":"dc21db7cf25bfe72"}},{"arxiv_id":"2308.13565","paper":"/paper/darwin-series-domain-specific-large-language","title":"DARWIN Series: Domain Specific Large Language Models for Natural Science","date":"2023-08-25","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"masterai-eam/darwin","path":"inference.py","file_url":"https://github.com/masterai-eam/darwin/blob/HEAD/inference.py","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"code_sha256_prefix":"6222762f51832899","mcp_get_code":{"code_sha256":"6222762f51832899"}},{"arxiv_id":"2305.12524","paper":"/paper/theoremqa-a-theorem-driven-question-answering","title":"TheoremQA: A Theorem-driven Question Answering dataset","date":"2023-05-21","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"THUDM/VisualGLM-6B","path":"model/chat.py","file_url":"https://github.com/THUDM/VisualGLM-6B/blob/HEAD/model/chat.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":"dc21db7cf25bfe72","mcp_get_code":{"code_sha256":"dc21db7cf25bfe72"}},{"arxiv_id":"2112.05682","paper":"/paper/self-attention-does-not-need-o-n-2-memory","title":"Self-attention Does Not Need $O(n^2)$ Memory","date":"2021-12-10","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"stability-ai/fastchat","path":"fastchat/model/model_chatglm.py","file_url":"https://github.com/stability-ai/fastchat/blob/HEAD/fastchat/model/model_chatglm.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":"dc21db7cf25bfe72","mcp_get_code":{"code_sha256":"dc21db7cf25bfe72"}},{"arxiv_id":"2024.findings-emnlp.559","paper":null,"title":"arXiv:2024.findings-emnlp.559","date":null,"month_inferred_from_arxiv_id":null,"title_source":null,"repo":"hitz-zentroa/cn-eval","path":"evaluation/JudgeLM-main/judgelm/model/chatglm_model.py","file_url":"https://github.com/hitz-zentroa/cn-eval/blob/HEAD/evaluation/JudgeLM-main/judgelm/model/chatglm_model.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":"dc21db7cf25bfe72","mcp_get_code":{"code_sha256":"dc21db7cf25bfe72"}}]}