{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:Y7KS2YN22NQU7VINAEUXTHR2TV","short_pith_number":"pith:Y7KS2YN2","canonical_record":{"source":{"id":"2505.19590","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-26T07:01:06Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"d21b544e2c7a446c1b6b4b60f6885a2ade25ecdf29f07a226c5c29afa33312ff","abstract_canon_sha256":"a83a896d73f839a3f15411ac5d5918d65098f6a0030ce7b0c3a81486cd90ea90"},"schema_version":"1.0"},"canonical_sha256":"c7d52d61bad3614fd50d0129799e3a9d6ff56e1d00ca57a05aa0c5c065316e4a","source":{"kind":"arxiv","id":"2505.19590","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.19590","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"arxiv_version","alias_value":"2505.19590v4","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.19590","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"pith_short_12","alias_value":"Y7KS2YN22NQU","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"Y7KS2YN22NQU7VIN","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"Y7KS2YN2","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:Y7KS2YN22NQU7VINAEUXTHR2TV","target":"record","payload":{"canonical_record":{"source":{"id":"2505.19590","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-26T07:01:06Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"d21b544e2c7a446c1b6b4b60f6885a2ade25ecdf29f07a226c5c29afa33312ff","abstract_canon_sha256":"a83a896d73f839a3f15411ac5d5918d65098f6a0030ce7b0c3a81486cd90ea90"},"schema_version":"1.0"},"canonical_sha256":"c7d52d61bad3614fd50d0129799e3a9d6ff56e1d00ca57a05aa0c5c065316e4a","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:50.140871Z","signature_b64":"fgz/tcPL4e/OckjF25z2BRJ+rHspf6b/m2pnczHk5xoY2ekIC+UTWUQ65COL6jSSoMpz7N0b0r2cx6Ukkc9gBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c7d52d61bad3614fd50d0129799e3a9d6ff56e1d00ca57a05aa0c5c065316e4a","last_reissued_at":"2026-05-17T23:38:50.140295Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:50.140295Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.19590","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"a+H6QHM1k8piSdkS7CSCXDLACMXWBREkT3TI7LInMzDABdoVfxpZX3+1kjm1ESN4QWo1nF3UG6V5NreLqGzaBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-02T22:20:02.445726Z"},"content_sha256":"0daebae2d012675d284d9a3e73bef70a0c77ed9057e3f9558423dbe1b6181812","schema_version":"1.0","event_id":"sha256:0daebae2d012675d284d9a3e73bef70a0c77ed9057e3f9558423dbe1b6181812"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:Y7KS2YN22NQU7VINAEUXTHR2TV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning to Reason without External Rewards","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Large language models can improve at reasoning by using only their own internal confidence as the reward signal.","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Aosong Feng, Dawn Song, Sergey Levine, Xuandong Zhao, Zhewei Kang","submitted_at":"2025-05-26T07:01:06Z","abstract_excerpt":"Training large language models (LLMs) for complex reasoning via Reinforcement Learning with Verifiable Rewards (RLVR) is effective but limited by reliance on costly, domain-specific supervision. We explore Reinforcement Learning from Internal Feedback (RLIF), a framework that enables LLMs to learn from intrinsic signals without external rewards or labeled data. We propose Intuitor, an RLIF method that uses a model's own confidence-termed self-certainty-as its sole reward signal. Intuitor replaces external rewards in Group Relative Policy Optimization (GRPO) with self-certainty scores, enabling"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments demonstrate that Intuitor matches GRPO's performance on mathematical benchmarks while achieving better generalization to out-of-domain tasks like code generation, without requiring gold solutions or test cases.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the model's self-certainty score reliably indicates correct reasoning and does not encourage reward hacking or overconfidence in incorrect outputs.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Intuitor enables LLMs to learn complex reasoning from self-certainty signals alone, matching supervised RL performance on math benchmarks while generalizing better to code generation without gold solutions or test cases.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Large language models can improve at reasoning by using only their own internal confidence as the reward signal.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"413e65f2c0095f14fbe0af388adc5b7212507fe4807fe16163e3629ff6355cda"},"source":{"id":"2505.19590","kind":"arxiv","version":4},"verdict":{"id":"7135c8df-9dc2-4fae-9f8f-bbd33c2ee816","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T21:13:34.177291Z","strongest_claim":"Experiments demonstrate that Intuitor matches GRPO's performance on mathematical benchmarks while achieving better generalization to out-of-domain tasks like code generation, without requiring gold solutions or test cases.","one_line_summary":"Intuitor enables LLMs to learn complex reasoning from self-certainty signals alone, matching supervised RL performance on math benchmarks while generalizing better to code generation without gold solutions or test cases.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the model's self-certainty score reliably indicates correct reasoning and does not encourage reward hacking or overconfidence in incorrect outputs.","pith_extraction_headline":"Large language models can improve at reasoning by using only their own internal confidence as the reward signal."},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":1,"snapshot_sha256":"bea6854910e03788df088a3ea5ad244a858f860c0aacd3932bfb93c950a30740"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"7135c8df-9dc2-4fae-9f8f-bbd33c2ee816"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"TphvCZE36er5/S1VvMTbMsPbmneBX8Q2TpYmuq6tUwVgYRUawHy/rbT3tRmKPdlUgJdMSd5H9P2d+/y7ttv7Cw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-02T22:20:02.446166Z"},"content_sha256":"dcf5a81ce5e40fc8e70c572fbb4bffc063de136605f308f290522e28b81ce902","schema_version":"1.0","event_id":"sha256:dcf5a81ce5e40fc8e70c572fbb4bffc063de136605f308f290522e28b81ce902"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/Y7KS2YN22NQU7VINAEUXTHR2TV/bundle.json","state_url":"https://pith.science/pith/Y7KS2YN22NQU7VINAEUXTHR2TV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/Y7KS2YN22NQU7VINAEUXTHR2TV/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-02T22:20:02Z","links":{"resolver":"https://pith.science/pith/Y7KS2YN22NQU7VINAEUXTHR2TV","bundle":"https://pith.science/pith/Y7KS2YN22NQU7VINAEUXTHR2TV/bundle.json","state":"https://pith.science/pith/Y7KS2YN22NQU7VINAEUXTHR2TV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/Y7KS2YN22NQU7VINAEUXTHR2TV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:Y7KS2YN22NQU7VINAEUXTHR2TV","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":"a83a896d73f839a3f15411ac5d5918d65098f6a0030ce7b0c3a81486cd90ea90","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-26T07:01:06Z","title_canon_sha256":"d21b544e2c7a446c1b6b4b60f6885a2ade25ecdf29f07a226c5c29afa33312ff"},"schema_version":"1.0","source":{"id":"2505.19590","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.19590","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"arxiv_version","alias_value":"2505.19590v4","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.19590","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"pith_short_12","alias_value":"Y7KS2YN22NQU","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"Y7KS2YN22NQU7VIN","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"Y7KS2YN2","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:dcf5a81ce5e40fc8e70c572fbb4bffc063de136605f308f290522e28b81ce902","target":"graph","created_at":"2026-05-17T23:38:50Z","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":"Experiments demonstrate that Intuitor matches GRPO's performance on mathematical benchmarks while achieving better generalization to out-of-domain tasks like code generation, without requiring gold solutions or test cases."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the model's self-certainty score reliably indicates correct reasoning and does not encourage reward hacking or overconfidence in incorrect outputs."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Intuitor enables LLMs to learn complex reasoning from self-certainty signals alone, matching supervised RL performance on math benchmarks while generalizing better to code generation without gold solutions or test cases."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Large language models can improve at reasoning by using only their own internal confidence as the reward signal."}],"snapshot_sha256":"413e65f2c0095f14fbe0af388adc5b7212507fe4807fe16163e3629ff6355cda"},"formal_canon":{"evidence_count":1,"snapshot_sha256":"bea6854910e03788df088a3ea5ad244a858f860c0aacd3932bfb93c950a30740"},"paper":{"abstract_excerpt":"Training large language models (LLMs) for complex reasoning via Reinforcement Learning with Verifiable Rewards (RLVR) is effective but limited by reliance on costly, domain-specific supervision. We explore Reinforcement Learning from Internal Feedback (RLIF), a framework that enables LLMs to learn from intrinsic signals without external rewards or labeled data. We propose Intuitor, an RLIF method that uses a model's own confidence-termed self-certainty-as its sole reward signal. Intuitor replaces external rewards in Group Relative Policy Optimization (GRPO) with self-certainty scores, enabling","authors_text":"Aosong Feng, Dawn Song, Sergey Levine, Xuandong Zhao, Zhewei Kang","cross_cats":["cs.CL"],"headline":"Large language models can improve at reasoning by using only their own internal confidence as the reward signal.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-26T07:01:06Z","title":"Learning to Reason without External Rewards"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.19590","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-15T21:13:34.177291Z","id":"7135c8df-9dc2-4fae-9f8f-bbd33c2ee816","model_set":{"reader":"grok-4.3"},"one_line_summary":"Intuitor enables LLMs to learn complex reasoning from self-certainty signals alone, matching supervised RL performance on math benchmarks while generalizing better to code generation without gold solutions or test cases.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Large language models can improve at reasoning by using only their own internal confidence as the reward signal.","strongest_claim":"Experiments demonstrate that Intuitor matches GRPO's performance on mathematical benchmarks while achieving better generalization to out-of-domain tasks like code generation, without requiring gold solutions or test cases.","weakest_assumption":"That the model's self-certainty score reliably indicates correct reasoning and does not encourage reward hacking or overconfidence in incorrect outputs."}},"verdict_id":"7135c8df-9dc2-4fae-9f8f-bbd33c2ee816"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:0daebae2d012675d284d9a3e73bef70a0c77ed9057e3f9558423dbe1b6181812","target":"record","created_at":"2026-05-17T23:38:50Z","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":"a83a896d73f839a3f15411ac5d5918d65098f6a0030ce7b0c3a81486cd90ea90","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-26T07:01:06Z","title_canon_sha256":"d21b544e2c7a446c1b6b4b60f6885a2ade25ecdf29f07a226c5c29afa33312ff"},"schema_version":"1.0","source":{"id":"2505.19590","kind":"arxiv","version":4}},"canonical_sha256":"c7d52d61bad3614fd50d0129799e3a9d6ff56e1d00ca57a05aa0c5c065316e4a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c7d52d61bad3614fd50d0129799e3a9d6ff56e1d00ca57a05aa0c5c065316e4a","first_computed_at":"2026-05-17T23:38:50.140295Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:50.140295Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"fgz/tcPL4e/OckjF25z2BRJ+rHspf6b/m2pnczHk5xoY2ekIC+UTWUQ65COL6jSSoMpz7N0b0r2cx6Ukkc9gBw==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:50.140871Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.19590","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0daebae2d012675d284d9a3e73bef70a0c77ed9057e3f9558423dbe1b6181812","sha256:dcf5a81ce5e40fc8e70c572fbb4bffc063de136605f308f290522e28b81ce902"],"state_sha256":"251bb7c2e73c7c0f7410371d91fd4cef694c0df9a50d4599b9b02f29dd7659c2"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Z1uzYrjqon+FPHufm14lqb72x6Y980zD1oWSV6hs9KGJPmG3LZalBCLXNuL9Z863MwQY7Li47QX+MtTiIuoGDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-02T22:20:02.450336Z","bundle_sha256":"11a6f4e842244781c0665da36c1e914edf05b3954bceef8015e82fee3b50f150"}}