{"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/causal-attention-bias","entry":"causal_attention_bias","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":17,"n_papers_ran":17,"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":2,"n_samples_ran":2,"n_samples_fingerprinted":0,"n_places":17,"n_places_pointer_only":5,"by_status":{"ran_honours":0,"ran_violates":0,"ran_draft_wrong":0,"ran_fixture":1,"ran":1,"unverified":0},"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":"2605.22297","paper":"/paper/arxiv-2605-22297","title":"One LR Doesn't Fit All: Heavy-Tail Guided Layerwise Learning Rates for LLMs","date":null,"month_inferred_from_arxiv_id":"2026-05","title_source":"syntology","repo":"hed-ucas/Layer-wise-Learning-Rate","path":"olmo/model.py","file_url":"https://github.com/hed-ucas/Layer-wise-Learning-Rate/blob/HEAD/olmo/model.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2605.16941","paper":"/paper/arxiv-2605-16941","title":"Roll Out and Roll Back: Diffusion LLMs are Their Own Efficiency Teachers","date":null,"month_inferred_from_arxiv_id":"2026-05","title_source":"syntology","repo":"Feng-Hong/WINO-DLLM","path":"LLaDA/modeling_llada.py","file_url":"https://github.com/Feng-Hong/WINO-DLLM/blob/HEAD/LLaDA/modeling_llada.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2604.21215","paper":"/paper/arxiv-2604-21215","title":"The Recurrent Transformer: Greater Effective Depth and Efficient Decoding","date":null,"month_inferred_from_arxiv_id":"2026-04","title_source":"syntology","repo":"geniucos/recurrent-transformer","path":"olmo/model.py","file_url":"https://github.com/geniucos/recurrent-transformer/blob/HEAD/olmo/model.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":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2604.15750","paper":"/paper/arxiv-2604-15750","title":"DepCap: Adaptive Block-Wise Parallel Decoding for Efficient Diffusion LM Inference","date":null,"month_inferred_from_arxiv_id":"2026-04","title_source":"syntology","repo":"X-Xia0828/DepCap","path":"llada/model/modeling_llada.py","file_url":"https://github.com/X-Xia0828/DepCap/blob/HEAD/llada/model/modeling_llada.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2602.11543","paper":"/paper/arxiv-2602-11543","title":"Pretraining A Large Language Model using Distributed GPUs: A Memory-Efficient Decentralized Paradigm","date":null,"month_inferred_from_arxiv_id":"2026-02","title_source":"syntology","repo":"zjr2000/SPES","path":"spes/model.py","file_url":"https://github.com/zjr2000/SPES/blob/HEAD/spes/model.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":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2602.02927","paper":"/paper/arxiv-2602-02927","title":"Training-Free Self-Correction for Multimodal Masked Diffusion Models","date":null,"month_inferred_from_arxiv_id":"2026-02","title_source":"syntology","repo":"huge123/FreeCorrection","path":"model/modeling_llada.py","file_url":"https://github.com/huge123/FreeCorrection/blob/HEAD/model/modeling_llada.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2602.02600","paper":"/paper/arxiv-2602-02600","title":"Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Models","date":"2026-02-01","month_inferred_from_arxiv_id":null,"title_source":"syntology","repo":"shuita2333/PAD-codes","path":"MMaDA-PAD/models/modeling_llada.py","file_url":"https://github.com/shuita2333/PAD-codes/blob/HEAD/MMaDA-PAD/models/modeling_llada.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2601.22527","paper":"/paper/arxiv-2601-22527","title":"ρ-EOS: Training-free Bidirectional Variable-Length Control for Masked Diffusion LLMs","date":null,"month_inferred_from_arxiv_id":"2026-01","title_source":"syntology","repo":"yjyddq/rho-EOS","path":"dllm_eval/models/modeling_llada.py","file_url":"https://github.com/yjyddq/rho-EOS/blob/HEAD/dllm_eval/models/modeling_llada.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":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2601.21484","paper":"/paper/arxiv-2601-21484","title":"ETS: Energy-Guided Test-Time Scaling for Training-Free RL Alignment","date":null,"month_inferred_from_arxiv_id":"2026-01","title_source":"syntology","repo":"sheriyuo/ETS","path":"llada/model/modeling_llada.py","file_url":"https://github.com/sheriyuo/ETS/blob/HEAD/llada/model/modeling_llada.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2601.02236","paper":"/paper/arxiv-2601-02236","title":"CD 4 LM: Consistency Distillation and aDaptive Decoding for Diffusion Language Models","date":null,"month_inferred_from_arxiv_id":"2026-01","title_source":"syntology","repo":"yihao-liang/CDLM","path":"src/model/modeling_llada.py","file_url":"https://github.com/yihao-liang/CDLM/blob/HEAD/src/model/modeling_llada.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":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2507.11097","paper":"/paper/the-devil-behind-the-mask-an-emergent-safety","title":"The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs","date":"2025-07-15","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"zichenwen1/dija","path":"MMaDA/models/modeling_llada.py","file_url":"https://github.com/zichenwen1/dija/blob/HEAD/MMaDA/models/modeling_llada.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":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2502.21309","paper":"/paper/2502-21309","title":"FANformer: Improving Large Language Models Through Effective Periodicity Modeling","date":"2025-02-28","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"YihongDong/FANformer","path":"olmo/model.py","file_url":"https://github.com/YihongDong/FANformer/blob/HEAD/olmo/model.py","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"8eb82bd106525d67","mcp_get_code":{"code_sha256":"8eb82bd106525d67"}},{"arxiv_id":"2501.00656","paper":"/paper/2-olmo-2-furious","title":"2 OLMo 2 Furious","date":"2024-12-31","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"allenai/olmo","path":"olmo/model.py","file_url":"https://github.com/allenai/olmo/blob/HEAD/olmo/model.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":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2411.12925","paper":"/paper/loss-to-loss-prediction-scaling-laws-for-all","title":"Loss-to-Loss Prediction: Scaling Laws for All Datasets","date":"2024-11-19","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"kempnerinstitute/loss-to-loss-olmo","path":"olmo/model.py","file_url":"https://github.com/kempnerinstitute/loss-to-loss-olmo/blob/HEAD/olmo/model.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":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2410.01380","paper":"/paper/knowledge-entropy-decay-during-language-model","title":"Knowledge Entropy Decay during Language Model Pretraining Hinders New Knowledge Acquisition","date":"2024-10-02","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"kaistAI/Knowledge-Entropy","path":"olmo/model.py","file_url":"https://github.com/kaistAI/Knowledge-Entropy/blob/HEAD/olmo/model.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":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2409.17146","paper":"/paper/molmo-and-pixmo-open-weights-and-open-data","title":"Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models","date":"2024-09-25","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"allenai/molmo","path":"olmo/model.py","file_url":"https://github.com/allenai/molmo/blob/HEAD/olmo/model.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":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}},{"arxiv_id":"2406.10670","paper":"/paper/color-filter-conditional-loss-reduction","title":"CoLoR-Filter: Conditional Loss Reduction Filtering for Targeted Language Model Pre-training","date":"2024-06-15","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"davidbrandfonbrener/color-filter-olmo","path":"olmo/model.py","file_url":"https://github.com/davidbrandfonbrener/color-filter-olmo/blob/HEAD/olmo/model.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":"f9abf6d74699cc1c","mcp_get_code":{"code_sha256":"f9abf6d74699cc1c"}}]}