{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:S2TJTMPPYHW32D5ESW4VJ672LC","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":"561e44b2432b4f4e957118a069dde7fae8fc9bfbf8aac408de3d4dfd1beed17d","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-17T18:18:32Z","title_canon_sha256":"4a2dd2ad87d2a6160674ba66eff6276562daa1a531df578df1f5f1e675a0628c"},"schema_version":"1.0","source":{"id":"2310.11511","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.11511","created_at":"2026-07-05T07:02:08Z"},{"alias_kind":"arxiv_version","alias_value":"2310.11511v1","created_at":"2026-07-05T07:02:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.11511","created_at":"2026-07-05T07:02:08Z"},{"alias_kind":"pith_short_12","alias_value":"S2TJTMPPYHW3","created_at":"2026-07-05T07:02:08Z"},{"alias_kind":"pith_short_16","alias_value":"S2TJTMPPYHW32D5E","created_at":"2026-07-05T07:02:08Z"},{"alias_kind":"pith_short_8","alias_value":"S2TJTMPP","created_at":"2026-07-05T07:02:08Z"}],"graph_snapshots":[{"event_id":"sha256:35c4eb16cf2570b2d05229104f5f70c4b9f5d9acf31b056c8c9f46e6fbfd6e75","target":"graph","created_at":"2026-07-05T07:02:08Z","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":"Self-RAG (7B and 13B parameters) significantly outperforms state-of-the-art LLMs and retrieval-augmented models on a diverse set of tasks. Specifically, Self-RAG outperforms ChatGPT and retrieval-augmented Llama2-chat on Open-domain QA, reasoning and fact verification tasks, and it shows significant gains in improving factuality and citation accuracy for long-form generations relative to these models."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That an arbitrary LM can be trained to generate and act on reflection tokens in a way that improves rather than degrades performance across tasks without introducing new failure modes from the added tokens."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Self-RAG trains LLMs to adaptively retrieve passages on demand and self-critique using reflection tokens, outperforming ChatGPT and retrieval-augmented Llama2 on QA, reasoning, and fact verification."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Self-RAG trains a single language model to adaptively retrieve passages on demand and critique its own outputs using special reflection tokens."}],"snapshot_sha256":"23d566505dfd6411625c27004b8d998868e212164cf9d9736afc9c8f84a533f2"},"formal_canon":{"evidence_count":3,"snapshot_sha256":"45cd1e37665bd876ce638c287082c7a51af5ea5aa951b2e3a37e679baf017c19"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2310.11511/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Despite their remarkable capabilities, large language models (LLMs) often produce responses containing factual inaccuracies due to their sole reliance on the parametric knowledge they encapsulate. Retrieval-Augmented Generation (RAG), an ad hoc approach that augments LMs with retrieval of relevant knowledge, decreases such issues. However, indiscriminately retrieving and incorporating a fixed number of retrieved passages, regardless of whether retrieval is necessary, or passages are relevant, diminishes LM versatility or can lead to unhelpful response generation. We introduce a new framework c","authors_text":"Akari Asai, Avirup Sil, Hannaneh Hajishirzi, Yizhong Wang, Zeqiu Wu","cross_cats":["cs.AI","cs.LG"],"headline":"Self-RAG trains a single language model to adaptively retrieve passages on demand and critique its own outputs using special reflection tokens.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-17T18:18:32Z","title":"Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection"},"references":{"count":154,"internal_anchors":23,"resolved_work":154,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Learning to retrieve reasoning paths over wikipedia graph for question answering","work_id":"2f4e6df6-c1e2-4f5e-80b4-e86b8e29dadd","year":2020},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Retrieval-based language models and applications","work_id":"8539632b-8089-43ed-b215-005ead5faa6e","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Task-aware retrieval with instructions","work_id":"31b1a479-fc74-4ee5-8290-a68ea7acc584","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":7,"title":"Flashattention: Fast and memory-efficient exact attention with io-awareness","work_id":"0a762bb5-5795-45ee-a82d-71bfb1e1f9a4","year":2022},{"cited_arxiv_id":"2309.11495","doi":"","is_internal_anchor":true,"ref_index":8,"title":"Chain-of-Verification Reduces Hallucination in Large Language Models","work_id":"e9e9983c-a72a-4864-9005-cd09bbd7876c","year":2023}],"snapshot_sha256":"57d13d5f94e8cf59559f8f5e742c799af4ccd15be3917592ec7e1650e834e232"},"source":{"id":"2310.11511","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-12T14:09:42.760870Z","id":"baa194fe-ec6e-45a6-bbd4-8745d6e57266","model_set":{"reader":"grok-4.3"},"one_line_summary":"Self-RAG trains LLMs to adaptively retrieve passages on demand and self-critique using reflection tokens, outperforming ChatGPT and retrieval-augmented Llama2 on QA, reasoning, and fact verification.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Self-RAG trains a single language model to adaptively retrieve passages on demand and critique its own outputs using special reflection tokens.","strongest_claim":"Self-RAG (7B and 13B parameters) significantly outperforms state-of-the-art LLMs and retrieval-augmented models on a diverse set of tasks. Specifically, Self-RAG outperforms ChatGPT and retrieval-augmented Llama2-chat on Open-domain QA, reasoning and fact verification tasks, and it shows significant gains in improving factuality and citation accuracy for long-form generations relative to these models.","weakest_assumption":"That an arbitrary LM can be trained to generate and act on reflection tokens in a way that improves rather than degrades performance across tasks without introducing new failure modes from the added tokens."}},"verdict_id":"baa194fe-ec6e-45a6-bbd4-8745d6e57266"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:79137f308a0419e04871c8e8a2c6a20048b269b1adb2fdde8c8f60bc9a6c4424","target":"record","created_at":"2026-07-05T07:02:08Z","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":"561e44b2432b4f4e957118a069dde7fae8fc9bfbf8aac408de3d4dfd1beed17d","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-17T18:18:32Z","title_canon_sha256":"4a2dd2ad87d2a6160674ba66eff6276562daa1a531df578df1f5f1e675a0628c"},"schema_version":"1.0","source":{"id":"2310.11511","kind":"arxiv","version":1}},"canonical_sha256":"96a699b1efc1edbd0fa495b954fbfa58bc4baefed86c2e351038dcb7ef6897de","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"96a699b1efc1edbd0fa495b954fbfa58bc4baefed86c2e351038dcb7ef6897de","first_computed_at":"2026-07-05T07:02:08.374924Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:02:08.374924Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KN1yHuMoUH4WIBNiSRsWEhAsg02IKbg+HmNY/wJzUMHqg6NdEswt6LtpHJnG/aJe4TnFwlIWiR5dFzSmLWU7Aw==","signature_status":"signed_v1","signed_at":"2026-07-05T07:02:08.375436Z","signed_message":"canonical_sha256_bytes"},"source_id":"2310.11511","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:79137f308a0419e04871c8e8a2c6a20048b269b1adb2fdde8c8f60bc9a6c4424","sha256:35c4eb16cf2570b2d05229104f5f70c4b9f5d9acf31b056c8c9f46e6fbfd6e75"],"state_sha256":"c6d40383e6a86b2941670b2ea809da402d00dc7a37192cb10fa92052f92cb9ba"}