{"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/confidence-interval","entry":"confidence_interval","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":11,"n_papers_ran":6,"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":5,"n_samples_fingerprinted":2,"n_places":11,"n_places_pointer_only":2,"by_status":{"ran_honours":0,"ran_violates":0,"ran_draft_wrong":0,"ran_fixture":0,"ran":5,"unverified":5},"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":"2609.04105","paper":"/paper/arxiv-2609-04105","title":"Hardware-Aware FP4 FlashAttention-4","date":null,"month_inferred_from_arxiv_id":"2026-09","title_source":"syntology","repo":"MrHuff/fp4-fa4","path":"results/fp4_fa4_reconstruction_20260805/build_summary.py","file_url":"https://github.com/MrHuff/fp4-fa4/blob/HEAD/results/fp4_fa4_reconstruction_20260805/build_summary.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":"f923592f03a7775c","mcp_get_code":{"code_sha256":"f923592f03a7775c"}},{"arxiv_id":"2605.11764","paper":"/paper/arxiv-2605-11764","title":"Decomposing the Generalization Gap in PROTAC Activity Prediction: Variance Attribution and the Inter-Laboratory Ceiling","date":null,"month_inferred_from_arxiv_id":"2026-05","title_source":"syntology","repo":"ThorKlm/PROTAC-Bench","path":"src/stats.py","file_url":"https://github.com/ThorKlm/PROTAC-Bench/blob/HEAD/src/stats.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"29cb7b75b3e65672","mcp_get_code":{"code_sha256":"29cb7b75b3e65672"}},{"arxiv_id":"2511.21140","paper":"/paper/arxiv-2511-21140","title":"How to Correctly Report LLM-as-a-Judge Evaluations","date":null,"month_inferred_from_arxiv_id":"2025-11","title_source":"syntology","repo":"UW-Madison-Lee-Lab/LLM-judge-reporting","path":"llm_judge_reporting/calibration.py","file_url":"https://github.com/UW-Madison-Lee-Lab/LLM-judge-reporting/blob/HEAD/llm_judge_reporting/calibration.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"8ae9f8d373851439","mcp_get_code":{"code_sha256":"8ae9f8d373851439"}},{"arxiv_id":"2505.21825","paper":"/paper/let-me-think-a-long-chain-of-thought-can-be","title":"Let Me Think! A Long Chain-of-Thought Can Be Worth Exponentially Many Short Ones","date":"2025-05-27","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"seyedparsa/let-me-think","path":"analysis/analysis_utils.py","file_url":"https://github.com/seyedparsa/let-me-think/blob/HEAD/analysis/analysis_utils.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"6f1fce08e48a8203","mcp_get_code":{"code_sha256":"6f1fce08e48a8203"}},{"arxiv_id":"2405.00218","paper":"/paper/constrained-decoding-for-secure-code","title":"Constrained Decoding for Secure Code Generation","date":"2024-04-30","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"dynamite321/codeguardplus","path":"correctness_eval.py","file_url":"https://github.com/dynamite321/codeguardplus/blob/HEAD/correctness_eval.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"4bd827a4fdf8d096","mcp_get_code":{"code_sha256":"4bd827a4fdf8d096"}},{"arxiv_id":"2404.04269","paper":"/paper/algorithmic-collective-action-in-recommender","title":"Algorithmic Collective Action in Recommender Systems: Promoting Songs by Reordering Playlists","date":"2024-03-19","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"joebaumann/recsys-collectiveaction","path":"recommender_system/2023_deezer_transformers/src/plot_results.py","file_url":"https://github.com/joebaumann/recsys-collectiveaction/blob/HEAD/recommender_system/2023_deezer_transformers/src/plot_results.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"105a23e9acb302d4","mcp_get_code":{"code_sha256":"105a23e9acb302d4"}},{"arxiv_id":"2310.18970","paper":"/paper/triage-characterizing-and-auditing-training","title":"TRIAGE: Characterizing and auditing training data for improved regression","date":"2023-10-29","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"seedatnabeel/triage","path":"src/utils.py","file_url":"https://github.com/seedatnabeel/triage/blob/HEAD/src/utils.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"23aa8c6c8dcfabf2","mcp_get_code":{"code_sha256":"23aa8c6c8dcfabf2"}},{"arxiv_id":"2307.06422","paper":null,"title":"arXiv:2307.06422","date":null,"month_inferred_from_arxiv_id":"2023-07","title_source":null,"repo":"thupchnsky/dp-gnn","path":"core/utils.py","file_url":"https://github.com/thupchnsky/dp-gnn/blob/HEAD/core/utils.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"f5e502a8909df4a3","mcp_get_code":{"code_sha256":"f5e502a8909df4a3"}},{"arxiv_id":"2203.00949","paper":"/paper/gap-differentially-private-graph-neural","title":"GAP: Differentially Private Graph Neural Networks with Aggregation Perturbation","date":"2022-03-02","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"sisaman/gap","path":"core/utils.py","file_url":"https://github.com/sisaman/gap/blob/HEAD/core/utils.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"CC0-1.0","inline_ok":true,"code_sha256_prefix":"f5e502a8909df4a3","mcp_get_code":{"code_sha256":"f5e502a8909df4a3"}},{"arxiv_id":"2201.13148","paper":"/paper/threshold-independent-evaluation-of-sound","title":"Threshold Independent Evaluation of Sound Event Detection Scores","date":"2022-01-31","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"fgnt/sed_scores_eval","path":"sed_scores_eval/base_modules/bootstrap.py","file_url":"https://github.com/fgnt/sed_scores_eval/blob/HEAD/sed_scores_eval/base_modules/bootstrap.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"e3b7b971e1627f6e","mcp_get_code":{"code_sha256":"e3b7b971e1627f6e"}},{"arxiv_id":"2111.02927","paper":"/paper/proximal-causal-inference-with-hidden","title":"Causal Inference with Hidden Mediators","date":"2021-11-04","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"syanga/hidmed","path":"hidmed/metrics.py","file_url":"https://github.com/syanga/hidmed/blob/HEAD/hidmed/metrics.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":"3fdbb6a18c9fa705","mcp_get_code":{"code_sha256":"3fdbb6a18c9fa705"}}]}