{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:6OXAXF5WGC227VSXW3QMQ3LYSR","short_pith_number":"pith:6OXAXF5W","canonical_record":{"source":{"id":"2502.01126","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-03T07:43:27Z","cross_cats_sorted":[],"title_canon_sha256":"b36f6b550e8e1a8fa4b3e42c0747e3f0f9854854041e8886d3688f1286b1b3a3","abstract_canon_sha256":"44668db419f0887b09fa985763d4e19ec2f95d4dee87987956e61bf134d196d2"},"schema_version":"1.0"},"canonical_sha256":"f3ae0b97b630b5afd657b6e0c86d78945804a4871a94adea803680b65b17cf2f","source":{"kind":"arxiv","id":"2502.01126","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.01126","created_at":"2026-07-05T10:08:43Z"},{"alias_kind":"arxiv_version","alias_value":"2502.01126v1","created_at":"2026-07-05T10:08:43Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.01126","created_at":"2026-07-05T10:08:43Z"},{"alias_kind":"pith_short_12","alias_value":"6OXAXF5WGC22","created_at":"2026-07-05T10:08:43Z"},{"alias_kind":"pith_short_16","alias_value":"6OXAXF5WGC227VSX","created_at":"2026-07-05T10:08:43Z"},{"alias_kind":"pith_short_8","alias_value":"6OXAXF5W","created_at":"2026-07-05T10:08:43Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:6OXAXF5WGC227VSXW3QMQ3LYSR","target":"record","payload":{"canonical_record":{"source":{"id":"2502.01126","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-03T07:43:27Z","cross_cats_sorted":[],"title_canon_sha256":"b36f6b550e8e1a8fa4b3e42c0747e3f0f9854854041e8886d3688f1286b1b3a3","abstract_canon_sha256":"44668db419f0887b09fa985763d4e19ec2f95d4dee87987956e61bf134d196d2"},"schema_version":"1.0"},"canonical_sha256":"f3ae0b97b630b5afd657b6e0c86d78945804a4871a94adea803680b65b17cf2f","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:08:43.162127Z","signature_b64":"OHwYBNonpWhIxsSE/AovQjN76oVi76Vq9WlP/CDk/M1cUHJ7WZOpYWgRloM7HmzpJMApEppm1VfSeeXKNw4OBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f3ae0b97b630b5afd657b6e0c86d78945804a4871a94adea803680b65b17cf2f","last_reissued_at":"2026-07-05T10:08:43.161692Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:08:43.161692Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.01126","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:08:43Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qUm9Xlq4NxmwhNURWm1w1uk5Nu8N0f7Iturw6izP5NSiFuosem9B2K7YstqqPYwG1PskaowDm/2VmcpqAm4KDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T20:13:54.219727Z"},"content_sha256":"410afd6fae02f8cbfbb0b061a17a3f7220939f0370b382ab9b4115d7f6a7b3a1","schema_version":"1.0","event_id":"sha256:410afd6fae02f8cbfbb0b061a17a3f7220939f0370b382ab9b4115d7f6a7b3a1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:6OXAXF5WGC227VSXW3QMQ3LYSR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Language Models Prefer What They Know: Relative Confidence Estimation via Confidence Preferences","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ananya Kumar, Percy Liang, Vaishnavi Shrivastava","submitted_at":"2025-02-03T07:43:27Z","abstract_excerpt":"Language models (LMs) should provide reliable confidence estimates to help users detect mistakes in their outputs and defer to human experts when necessary. Asking a language model to assess its confidence (\"Score your confidence from 0-1.\") is a natural way of evaluating its uncertainty. However, models struggle to provide absolute assessments of confidence (i.e. judging confidence in answering a question independent of other questions) and the coarse-grained scores they produce are not useful for evaluating the correctness of their answers. We propose relative confidence estimation, where we"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.01126","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2502.01126/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:08:43Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"imzExlLjQ6eU/RVQjvIsEsuHzlVZq2PLATeA7jtXm0Do0uT+EPFbCdmumhjRBG/QU7sU1LGqdK9fd+iii6c5Aw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T20:13:54.220284Z"},"content_sha256":"72f0dc741910b1df6344ad730b4302cf5738d6174b567b1d8babb83b75765eca","schema_version":"1.0","event_id":"sha256:72f0dc741910b1df6344ad730b4302cf5738d6174b567b1d8babb83b75765eca"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/6OXAXF5WGC227VSXW3QMQ3LYSR/bundle.json","state_url":"https://pith.science/pith/6OXAXF5WGC227VSXW3QMQ3LYSR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/6OXAXF5WGC227VSXW3QMQ3LYSR/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":"2026-08-09T20:13:54Z","links":{"resolver":"https://pith.science/pith/6OXAXF5WGC227VSXW3QMQ3LYSR","bundle":"https://pith.science/pith/6OXAXF5WGC227VSXW3QMQ3LYSR/bundle.json","state":"https://pith.science/pith/6OXAXF5WGC227VSXW3QMQ3LYSR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/6OXAXF5WGC227VSXW3QMQ3LYSR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:6OXAXF5WGC227VSXW3QMQ3LYSR","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":"44668db419f0887b09fa985763d4e19ec2f95d4dee87987956e61bf134d196d2","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-03T07:43:27Z","title_canon_sha256":"b36f6b550e8e1a8fa4b3e42c0747e3f0f9854854041e8886d3688f1286b1b3a3"},"schema_version":"1.0","source":{"id":"2502.01126","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.01126","created_at":"2026-07-05T10:08:43Z"},{"alias_kind":"arxiv_version","alias_value":"2502.01126v1","created_at":"2026-07-05T10:08:43Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.01126","created_at":"2026-07-05T10:08:43Z"},{"alias_kind":"pith_short_12","alias_value":"6OXAXF5WGC22","created_at":"2026-07-05T10:08:43Z"},{"alias_kind":"pith_short_16","alias_value":"6OXAXF5WGC227VSX","created_at":"2026-07-05T10:08:43Z"},{"alias_kind":"pith_short_8","alias_value":"6OXAXF5W","created_at":"2026-07-05T10:08:43Z"}],"graph_snapshots":[{"event_id":"sha256:72f0dc741910b1df6344ad730b4302cf5738d6174b567b1d8babb83b75765eca","target":"graph","created_at":"2026-07-05T10:08:43Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2502.01126/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Language models (LMs) should provide reliable confidence estimates to help users detect mistakes in their outputs and defer to human experts when necessary. Asking a language model to assess its confidence (\"Score your confidence from 0-1.\") is a natural way of evaluating its uncertainty. However, models struggle to provide absolute assessments of confidence (i.e. judging confidence in answering a question independent of other questions) and the coarse-grained scores they produce are not useful for evaluating the correctness of their answers. We propose relative confidence estimation, where we","authors_text":"Ananya Kumar, Percy Liang, Vaishnavi Shrivastava","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-03T07:43:27Z","title":"Language Models Prefer What They Know: Relative Confidence Estimation via Confidence Preferences"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.01126","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:410afd6fae02f8cbfbb0b061a17a3f7220939f0370b382ab9b4115d7f6a7b3a1","target":"record","created_at":"2026-07-05T10:08:43Z","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":"44668db419f0887b09fa985763d4e19ec2f95d4dee87987956e61bf134d196d2","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-03T07:43:27Z","title_canon_sha256":"b36f6b550e8e1a8fa4b3e42c0747e3f0f9854854041e8886d3688f1286b1b3a3"},"schema_version":"1.0","source":{"id":"2502.01126","kind":"arxiv","version":1}},"canonical_sha256":"f3ae0b97b630b5afd657b6e0c86d78945804a4871a94adea803680b65b17cf2f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f3ae0b97b630b5afd657b6e0c86d78945804a4871a94adea803680b65b17cf2f","first_computed_at":"2026-07-05T10:08:43.161692Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:08:43.161692Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"OHwYBNonpWhIxsSE/AovQjN76oVi76Vq9WlP/CDk/M1cUHJ7WZOpYWgRloM7HmzpJMApEppm1VfSeeXKNw4OBg==","signature_status":"signed_v1","signed_at":"2026-07-05T10:08:43.162127Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.01126","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:410afd6fae02f8cbfbb0b061a17a3f7220939f0370b382ab9b4115d7f6a7b3a1","sha256:72f0dc741910b1df6344ad730b4302cf5738d6174b567b1d8babb83b75765eca"],"state_sha256":"1a60402ac09212dbe2140d8006755f6e185c3f738ec80377e138f1b6ac89c72e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ztb1Zpo1FILIDdyLNs5U99sDU3lTwiNqFNlc9rqylwLKupqk4QN5KOQddRgsSFPuOaXyct1H4yJCBcfSJiPzAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T20:13:54.223807Z","bundle_sha256":"08b9a0af325557bd882d55e809e82073d9dc5c00281b5fda0993370f5b50d499"}}