{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:2TRLZM4RVJI4V2U6RCU7O7TRJA","short_pith_number":"pith:2TRLZM4R","canonical_record":{"source":{"id":"2505.21011","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-27T10:45:29Z","cross_cats_sorted":[],"title_canon_sha256":"c50af4fa4c352c53b7a63f71edc00698ec262aca1b8bfe6d30a13c83a961e210","abstract_canon_sha256":"fffd324a585711da866cea60efbb1dedf23e53a504464de14b2284d5ce623240"},"schema_version":"1.0"},"canonical_sha256":"d4e2bcb391aa51caea9e88a9f77e71480389a94aa4bfb486e85bf7b44a5a80f4","source":{"kind":"arxiv","id":"2505.21011","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.21011","created_at":"2026-07-05T11:10:22Z"},{"alias_kind":"arxiv_version","alias_value":"2505.21011v1","created_at":"2026-07-05T11:10:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.21011","created_at":"2026-07-05T11:10:22Z"},{"alias_kind":"pith_short_12","alias_value":"2TRLZM4RVJI4","created_at":"2026-07-05T11:10:22Z"},{"alias_kind":"pith_short_16","alias_value":"2TRLZM4RVJI4V2U6","created_at":"2026-07-05T11:10:22Z"},{"alias_kind":"pith_short_8","alias_value":"2TRLZM4R","created_at":"2026-07-05T11:10:22Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:2TRLZM4RVJI4V2U6RCU7O7TRJA","target":"record","payload":{"canonical_record":{"source":{"id":"2505.21011","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-27T10:45:29Z","cross_cats_sorted":[],"title_canon_sha256":"c50af4fa4c352c53b7a63f71edc00698ec262aca1b8bfe6d30a13c83a961e210","abstract_canon_sha256":"fffd324a585711da866cea60efbb1dedf23e53a504464de14b2284d5ce623240"},"schema_version":"1.0"},"canonical_sha256":"d4e2bcb391aa51caea9e88a9f77e71480389a94aa4bfb486e85bf7b44a5a80f4","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:10:22.381693Z","signature_b64":"feSFvzkD73y9UMkqiAnTcBtqlg2JcwjFGVvq+9EpAH10WJPgYq/ktER2ecF86WF0fhSqhiy3Nf+i2s4n+MF5Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d4e2bcb391aa51caea9e88a9f77e71480389a94aa4bfb486e85bf7b44a5a80f4","last_reissued_at":"2026-07-05T11:10:22.381252Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:10:22.381252Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.21011","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":"2026-07-05T11:10:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"BpmKn67jRWmnbrr344HLAsPG9FXfqfuYYftinvy3gVVXryT+I+KNY2MMAIhWMkiePnCmYdJVN443HdV6PvcsAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T11:45:50.444347Z"},"content_sha256":"b91153b077a48e775e3d591b27d91b661be7e008cdbc59fbb7a91dae5c464da6","schema_version":"1.0","event_id":"sha256:b91153b077a48e775e3d591b27d91b661be7e008cdbc59fbb7a91dae5c464da6"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:2TRLZM4RVJI4V2U6RCU7O7TRJA","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"LLMs are Frequency Pattern Learners in Natural Language Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Liang Cheng, Mark Steedman, Zhaowei Wang","submitted_at":"2025-05-27T10:45:29Z","abstract_excerpt":"While fine-tuning LLMs on NLI corpora improves their inferential performance, the underlying mechanisms driving this improvement remain largely opaque. In this work, we conduct a series of experiments to investigate what LLMs actually learn during fine-tuning. We begin by analyzing predicate frequencies in premises and hypotheses across NLI datasets and identify a consistent frequency bias, where predicates in hypotheses occur more frequently than those in premises for positive instances. To assess the impact of this bias, we evaluate both standard and NLI fine-tuned LLMs on bias-consistent an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.21011","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/2505.21011/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":"2026-07-05T11:10:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZAhOR5zBD9N/yeC0ZhpiXbMcS+nZW63//tln8oUKIP4eiCYvNAyhMdVrSxdgwy5a3VCp8a2fSNFC1R0cp7NjDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T11:45:50.444904Z"},"content_sha256":"858a7ba96e0df9e1713a4c788898447f1c597499750938091c7c5958479e3109","schema_version":"1.0","event_id":"sha256:858a7ba96e0df9e1713a4c788898447f1c597499750938091c7c5958479e3109"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/2TRLZM4RVJI4V2U6RCU7O7TRJA/bundle.json","state_url":"https://pith.science/pith/2TRLZM4RVJI4V2U6RCU7O7TRJA/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/2TRLZM4RVJI4V2U6RCU7O7TRJA/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":"2026-08-09T11:45:50Z","links":{"resolver":"https://pith.science/pith/2TRLZM4RVJI4V2U6RCU7O7TRJA","bundle":"https://pith.science/pith/2TRLZM4RVJI4V2U6RCU7O7TRJA/bundle.json","state":"https://pith.science/pith/2TRLZM4RVJI4V2U6RCU7O7TRJA/state.json","well_known_bundle":"https://pith.science/.well-known/pith/2TRLZM4RVJI4V2U6RCU7O7TRJA/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:2TRLZM4RVJI4V2U6RCU7O7TRJA","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":"fffd324a585711da866cea60efbb1dedf23e53a504464de14b2284d5ce623240","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-27T10:45:29Z","title_canon_sha256":"c50af4fa4c352c53b7a63f71edc00698ec262aca1b8bfe6d30a13c83a961e210"},"schema_version":"1.0","source":{"id":"2505.21011","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.21011","created_at":"2026-07-05T11:10:22Z"},{"alias_kind":"arxiv_version","alias_value":"2505.21011v1","created_at":"2026-07-05T11:10:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.21011","created_at":"2026-07-05T11:10:22Z"},{"alias_kind":"pith_short_12","alias_value":"2TRLZM4RVJI4","created_at":"2026-07-05T11:10:22Z"},{"alias_kind":"pith_short_16","alias_value":"2TRLZM4RVJI4V2U6","created_at":"2026-07-05T11:10:22Z"},{"alias_kind":"pith_short_8","alias_value":"2TRLZM4R","created_at":"2026-07-05T11:10:22Z"}],"graph_snapshots":[{"event_id":"sha256:858a7ba96e0df9e1713a4c788898447f1c597499750938091c7c5958479e3109","target":"graph","created_at":"2026-07-05T11:10:22Z","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/2505.21011/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"While fine-tuning LLMs on NLI corpora improves their inferential performance, the underlying mechanisms driving this improvement remain largely opaque. In this work, we conduct a series of experiments to investigate what LLMs actually learn during fine-tuning. We begin by analyzing predicate frequencies in premises and hypotheses across NLI datasets and identify a consistent frequency bias, where predicates in hypotheses occur more frequently than those in premises for positive instances. To assess the impact of this bias, we evaluate both standard and NLI fine-tuned LLMs on bias-consistent an","authors_text":"Liang Cheng, Mark Steedman, Zhaowei Wang","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-27T10:45:29Z","title":"LLMs are Frequency Pattern Learners in Natural Language Inference"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.21011","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:b91153b077a48e775e3d591b27d91b661be7e008cdbc59fbb7a91dae5c464da6","target":"record","created_at":"2026-07-05T11:10:22Z","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":"fffd324a585711da866cea60efbb1dedf23e53a504464de14b2284d5ce623240","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-27T10:45:29Z","title_canon_sha256":"c50af4fa4c352c53b7a63f71edc00698ec262aca1b8bfe6d30a13c83a961e210"},"schema_version":"1.0","source":{"id":"2505.21011","kind":"arxiv","version":1}},"canonical_sha256":"d4e2bcb391aa51caea9e88a9f77e71480389a94aa4bfb486e85bf7b44a5a80f4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d4e2bcb391aa51caea9e88a9f77e71480389a94aa4bfb486e85bf7b44a5a80f4","first_computed_at":"2026-07-05T11:10:22.381252Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:10:22.381252Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"feSFvzkD73y9UMkqiAnTcBtqlg2JcwjFGVvq+9EpAH10WJPgYq/ktER2ecF86WF0fhSqhiy3Nf+i2s4n+MF5Bw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:10:22.381693Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.21011","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b91153b077a48e775e3d591b27d91b661be7e008cdbc59fbb7a91dae5c464da6","sha256:858a7ba96e0df9e1713a4c788898447f1c597499750938091c7c5958479e3109"],"state_sha256":"f34312b8015b3731fc72b07f4458aac2dda2186097bef52f4c81b5e9613d0d29"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2Vdfd5bTXIGc0gGmShZGg+48cl+wxaDj7EUN+Prwq0VI+h+oYEwh7aC6zZHs76QdvQtU6jEWTn8i+XWlMP6mCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T11:45:50.450416Z","bundle_sha256":"4d721fc79cb35f02b6ab61036e87676179575cd7ed8a2c9a2a1cb5b633961736"}}