{"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/extract-step","entry":"extract_step","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":22,"n_papers_ran":21,"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":3,"n_samples_ran":2,"n_samples_fingerprinted":2,"n_places":22,"n_places_pointer_only":7,"by_status":{"ran_honours":2,"ran_violates":0,"ran_draft_wrong":0,"ran_fixture":0,"ran":0,"unverified":1},"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":"2606.10646","paper":"/paper/arxiv-2606-10646","title":"How Does Reasoning Flow? Tracing Attention-Induced Information Flow for Targeted RL in LLMs","date":null,"month_inferred_from_arxiv_id":"2026-06","title_source":"syntology","repo":"Jiayi-Pan/TinyZero","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/Jiayi-Pan/TinyZero/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2606.07705","paper":"/paper/arxiv-2606-07705","title":"SAW: Stage-Aware Dynamic Weighting for Multi-Objective Reinforcement Learning in Large Language Models","date":null,"month_inferred_from_arxiv_id":"2026-06","title_source":"syntology","repo":"Zhaolutuan/SAW","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/Zhaolutuan/SAW/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2602.15564","paper":"/paper/arxiv-2602-15564","title":"Beyond Static Pipelines: Learning Dynamic Workflows for Text-to-SQL","date":null,"month_inferred_from_arxiv_id":"2026-02","title_source":"syntology","repo":"Satissss/SquRL","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/Satissss/SquRL/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2602.12735","paper":"/paper/arxiv-2602-12735","title":"VimRAG: Navigating Massive Visual Context in Retrieval-Augmented Generation via Multimodal Memory Graph","date":null,"month_inferred_from_arxiv_id":"2026-02","title_source":"syntology","repo":"Alibaba-NLP/VRAG","path":"VRAG-RL/verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/Alibaba-NLP/VRAG/blob/HEAD/VRAG-RL/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2602.12222","paper":"/paper/arxiv-2602-12222","title":"Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training","date":"2026-02-12","month_inferred_from_arxiv_id":null,"title_source":"syntology","repo":"zhangmiaosen2000/Towards-On-Policy-SFT","path":"IDFT/verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/zhangmiaosen2000/Towards-On-Policy-SFT/blob/HEAD/IDFT/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2602.04284","paper":"/paper/arxiv-2602-04284","title":"Agent-Omit: Adaptive Context Omission for Efficient LLM Agents","date":null,"month_inferred_from_arxiv_id":"2026-02","title_source":"syntology","repo":"usail-hkust/Agent-Omit","path":"AgentOmit-RL/verl/agent_trainer/fsdp_sft_trainer.py","file_url":"https://github.com/usail-hkust/Agent-Omit/blob/HEAD/AgentOmit-RL/verl/agent_trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2601.09233","paper":"/paper/arxiv-2601-09233","title":"GIFT: Reconciling Post-Training Objectives via Finite-Temperature Gibbs Initialization","date":null,"month_inferred_from_arxiv_id":"2026-01","title_source":"syntology","repo":"zzy1127/GIFT","path":"verl/verl/trainer/fsdp_gift_trainer.py","file_url":"https://github.com/zzy1127/GIFT/blob/HEAD/verl/verl/trainer/fsdp_gift_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2511.00405","paper":"/paper/arxiv-2511-00405","title":"UME-R1: Exploring Reasoning-Driven Generative Multimodal Embeddings","date":null,"month_inferred_from_arxiv_id":"2025-11","title_source":"syntology","repo":"XMUDeepLIT/UME-R1","path":"src/eval/VLM2Vec/adhoc/gather_score_byckpt_aws.py","file_url":"https://github.com/XMUDeepLIT/UME-R1/blob/HEAD/src/eval/VLM2Vec/adhoc/gather_score_byckpt_aws.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":"981eb19d513b4f7e","mcp_get_code":{"code_sha256":"981eb19d513b4f7e"}},{"arxiv_id":"2505.24630","paper":"/paper/the-hallucination-dilemma-factuality-aware","title":"The Hallucination Dilemma: Factuality-Aware Reinforcement Learning for Large Reasoning Models","date":"2025-05-30","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":null,"path":"","file_url":null,"status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":null,"inline_ok":false,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2505.18917","paper":"/paper/behavior-injection-preparing-language-models","title":"Behavior Injection: Preparing Language Models for Reinforcement Learning","date":"2025-05-25","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"czp16/bridge-llm-reasoning","path":"simple_verl/sverl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/czp16/bridge-llm-reasoning/blob/HEAD/simple_verl/sverl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2505.18298","paper":"/paper/thinking-fast-and-right-balancing-accuracy","title":"Thinking Fast and Right: Balancing Accuracy and Reasoning Length with Adaptive Rewards","date":"2025-05-23","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":null,"path":"","file_url":null,"status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":null,"inline_ok":false,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2505.16984","paper":"/paper/uft-unifying-supervised-and-reinforcement","title":"UFT: Unifying Supervised and Reinforcement Fine-Tuning","date":"2025-05-22","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"liumy2010/uft","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/liumy2010/uft/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2505.12929","paper":"/paper/do-not-let-low-probability-tokens-over","title":"Do Not Let Low-Probability Tokens Over-Dominate in RL for LLMs","date":"2025-05-19","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"zhyang2226/ar-lopti","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/zhyang2226/ar-lopti/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2505.12366","paper":"/paper/disco-reinforcing-large-reasoning-models-with","title":"DisCO: Reinforcing Large Reasoning Models with Discriminative Constrained Optimization","date":"2025-05-18","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"optimization-ai/disco","path":"verl/verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/optimization-ai/disco/blob/HEAD/verl/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2505.04588","paper":"/paper/zerosearch-incentivize-the-search-capability","title":"ZeroSearch: Incentivize the Search Capability of LLMs without Searching","date":"2025-05-07","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"alibaba-nlp/zerosearch","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/alibaba-nlp/zerosearch/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2504.05520","paper":"/paper/efficient-reinforcement-finetuning-via","title":"Efficient Reinforcement Finetuning via Adaptive Curriculum Learning","date":"2025-04-07","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"uscnlp-lime/verl","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/uscnlp-lime/verl/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2504.01698","paper":"/paper/tom-rl-reinforcement-learning-unlocks-theory","title":"Do Theory of Mind Benchmarks Need Explicit Human-like Reasoning in Language Models?","date":"2025-04-02","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"bigai-ai/ToM-RL","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/bigai-ai/ToM-RL/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2502.14768","paper":"/paper/logic-rl-unleashing-llm-reasoning-with-rule","title":"Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learning","date":"2025-02-20","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"Unakar/Logic-RL","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/Unakar/Logic-RL/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2410.21236","paper":"/paper/flaming-hot-initiation-with-regular-execution","title":"Flaming-hot Initiation with Regular Execution Sampling for Large Language Models","date":"2024-10-28","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"yaof20/real","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/yaof20/real/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2408.09027","paper":"/paper/efficient-autoregressive-audio-modeling-via","title":"Efficient Autoregressive Audio Modeling via Next-Scale Prediction","date":"2024-08-16","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"qiuk2/aar","path":"train_AAR_mpi.py","file_url":"https://github.com/qiuk2/aar/blob/HEAD/train_AAR_mpi.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"d09dbeb0d659afef","mcp_get_code":{"code_sha256":"d09dbeb0d659afef"}},{"arxiv_id":"openreview_QrC8OgQyOI","paper":null,"title":"arXiv:openreview_QrC8OgQyOI","date":null,"month_inferred_from_arxiv_id":null,"title_source":null,"repo":"viiika/Prism","path":"Dream/Dream_Prism/src/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/viiika/Prism/blob/HEAD/Dream/Dream_Prism/src/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}},{"arxiv_id":"2025.findings-emnlp.1015","paper":null,"title":"arXiv:2025.findings-emnlp.1015","date":null,"month_inferred_from_arxiv_id":null,"title_source":null,"repo":"fzp0424/MT-R1-Zero","path":"verl/trainer/fsdp_sft_trainer.py","file_url":"https://github.com/fzp0424/MT-R1-Zero/blob/HEAD/verl/trainer/fsdp_sft_trainer.py","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"216e28c040173a61","mcp_get_code":{"code_sha256":"216e28c040173a61"}}]}