{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:656DEUNACXYL43CYGYJA4Y6JXK","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"536774355440d16d32852d1d5b7dbfec57261204744c37a7e2a685e9e8d8360d","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-30T15:25:45Z","title_canon_sha256":"314e98a7f73c2b0c1bc6afb2a80a822404b6a592830fef72602665def411d52f"},"schema_version":"1.0","source":{"id":"2305.19118","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.19118","created_at":"2026-05-18T03:30:52Z"},{"alias_kind":"arxiv_version","alias_value":"2305.19118v4","created_at":"2026-05-18T03:30:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.19118","created_at":"2026-05-18T03:30:52Z"},{"alias_kind":"pith_short_12","alias_value":"656DEUNACXYL","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"656DEUNACXYL43CY","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"656DEUNA","created_at":"2026-05-18T12:33:33Z"}],"graph_snapshots":[{"event_id":"sha256:d5e4cd28086434007ddc648e53a42395ab3da889471b2614ba9f2c8df2b285b8","target":"graph","created_at":"2026-05-18T03:30:52Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Our MAD framework encourages divergent thinking in LLMs which would be helpful for tasks that require deep levels of contemplation. Experiment results on two challenging datasets, commonsense machine translation and counter-intuitive arithmetic reasoning, demonstrate the effectiveness of our MAD framework."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That a judge LLM can fairly evaluate and synthesize the debate without inheriting the same Degeneration-of-Thought bias, and that the tit-for-tat dynamic reliably produces novel thoughts rather than escalating errors."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Multi-agent debate with tit-for-tat arguments and a judge LLM improves reasoning by preventing LLMs from locking into incorrect initial solutions."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Large language models overcome stuck reasoning by having multiple agents argue tit-for-tat under a judge instead of reflecting alone."}],"snapshot_sha256":"085674e93ac54155e332f700862953f255318eb6d06e2bb45bb3b7918b72a98c"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"45898dc983777a68eb914b3df3c7add05f2ade743dfddc4c91658651bc751620"},"paper":{"abstract_excerpt":"Modern large language models (LLMs) like ChatGPT have shown remarkable performance on general language tasks but still struggle on complex reasoning tasks, which drives the research on cognitive behaviors of LLMs to explore human-like problem-solving strategies. Along this direction, one representative strategy is self-reflection, which asks an LLM to refine the solution with the feedback generated by itself iteratively. However, our study shows that such reflection-style methods suffer from the Degeneration-of-Thought (DoT) problem: once the LLM has established confidence in its solutions, it","authors_text":"Rui Wang, Shuming Shi, Tian Liang, Wenxiang Jiao, Xing Wang, Yan Wang, Yujiu Yang, Zhaopeng Tu, Zhiwei He","cross_cats":[],"headline":"Large language models overcome stuck reasoning by having multiple agents argue tit-for-tat under a judge instead of reflecting alone.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-30T15:25:45Z","title":"Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate"},"references":{"count":286,"internal_anchors":10,"resolved_work":286,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing , pages=","work_id":"2bd92650-b903-4f69-9ae1-8c4085db98a8","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"A Multitask, Multilingual, Multimodal Evaluation of ChatGPT on Reasoning, Hallucination, and Interactivity , author=. Proceedings of the 13th International Joint Conference on Natural Language Process","work_id":"ceda3664-76be-4173-ac42-7b31fdf7124c","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":7,"title":"Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing , pages=","work_id":"2232ea1d-9981-4dc7-8380-6b7c6e81f564","year":2015},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":12,"title":"Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology , pages=","work_id":"cf0f4dfd-48c8-4bea-b512-f8814deee1fd","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":13,"title":"Advances in neural information processing systems , volume=","work_id":"f3923cef-fa32-49c8-8280-9c6bb5d18d27","year":null}],"snapshot_sha256":"856cc4182993b9ff6f11949be4c4f8fdff74f8f1d028a55fd563a3e20561d716"},"source":{"id":"2305.19118","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-13T23:56:00.948856Z","id":"dfd64283-d483-4049-bf8e-c6e99c890a83","model_set":{"reader":"grok-4.3"},"one_line_summary":"Multi-agent debate with tit-for-tat arguments and a judge LLM improves reasoning by preventing LLMs from locking into incorrect initial solutions.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Large language models overcome stuck reasoning by having multiple agents argue tit-for-tat under a judge instead of reflecting alone.","strongest_claim":"Our MAD framework encourages divergent thinking in LLMs which would be helpful for tasks that require deep levels of contemplation. Experiment results on two challenging datasets, commonsense machine translation and counter-intuitive arithmetic reasoning, demonstrate the effectiveness of our MAD framework.","weakest_assumption":"That a judge LLM can fairly evaluate and synthesize the debate without inheriting the same Degeneration-of-Thought bias, and that the tit-for-tat dynamic reliably produces novel thoughts rather than escalating errors."}},"verdict_id":"dfd64283-d483-4049-bf8e-c6e99c890a83"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:266f5c04bd9372c2934b0400b614b34998d5f10d0e516a1789bfbfc37be66e27","target":"record","created_at":"2026-05-18T03:30:52Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"536774355440d16d32852d1d5b7dbfec57261204744c37a7e2a685e9e8d8360d","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-30T15:25:45Z","title_canon_sha256":"314e98a7f73c2b0c1bc6afb2a80a822404b6a592830fef72602665def411d52f"},"schema_version":"1.0","source":{"id":"2305.19118","kind":"arxiv","version":4}},"canonical_sha256":"f77c3251a015f0be6c5836120e63c9ba9c42f0af9f480f7a1fb7885929e16031","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f77c3251a015f0be6c5836120e63c9ba9c42f0af9f480f7a1fb7885929e16031","first_computed_at":"2026-05-18T03:30:52.354097Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T03:30:52.354097Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"yV38JiDuAZnAgLkNkNuriNZI+eXHgf0jLx9sfXIMUZcF0djnuC5yj2XzUrfroez2kYw+trCDf7UJkABzsKHcBg==","signature_status":"signed_v1","signed_at":"2026-05-18T03:30:52.355108Z","signed_message":"canonical_sha256_bytes"},"source_id":"2305.19118","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:266f5c04bd9372c2934b0400b614b34998d5f10d0e516a1789bfbfc37be66e27","sha256:d5e4cd28086434007ddc648e53a42395ab3da889471b2614ba9f2c8df2b285b8"],"state_sha256":"6f9f4eb3f8311af1ed5a8f5ff8d15d70d378e2eace0d401409aa7b8e4ad9651f"}