{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:4I52FZU52Q3JVQJGK3KJI4CHIF","short_pith_number":"pith:4I52FZU5","canonical_record":{"source":{"id":"2204.08039","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-17T15:55:18Z","cross_cats_sorted":[],"title_canon_sha256":"1bdf236dc133ada1fba614dba8438fd6a9800a279f06fccca3d4b5ee61ee3812","abstract_canon_sha256":"e7a20bccdca2b181bb8c23220607ee46b21a413e60657615fe2c9556cac11e76"},"schema_version":"1.0"},"canonical_sha256":"e23ba2e69dd4369ac12656d4947047414c6d69bd0cbcad7e02a6d00a76156307","source":{"kind":"arxiv","id":"2204.08039","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2204.08039","created_at":"2026-07-05T04:15:21Z"},{"alias_kind":"arxiv_version","alias_value":"2204.08039v1","created_at":"2026-07-05T04:15:21Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.08039","created_at":"2026-07-05T04:15:21Z"},{"alias_kind":"pith_short_12","alias_value":"4I52FZU52Q3J","created_at":"2026-07-05T04:15:21Z"},{"alias_kind":"pith_short_16","alias_value":"4I52FZU52Q3JVQJG","created_at":"2026-07-05T04:15:21Z"},{"alias_kind":"pith_short_8","alias_value":"4I52FZU5","created_at":"2026-07-05T04:15:21Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:4I52FZU52Q3JVQJGK3KJI4CHIF","target":"record","payload":{"canonical_record":{"source":{"id":"2204.08039","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-17T15:55:18Z","cross_cats_sorted":[],"title_canon_sha256":"1bdf236dc133ada1fba614dba8438fd6a9800a279f06fccca3d4b5ee61ee3812","abstract_canon_sha256":"e7a20bccdca2b181bb8c23220607ee46b21a413e60657615fe2c9556cac11e76"},"schema_version":"1.0"},"canonical_sha256":"e23ba2e69dd4369ac12656d4947047414c6d69bd0cbcad7e02a6d00a76156307","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:15:21.411297Z","signature_b64":"MzIvtiCzZnkuwgmeEaSjq0q9RFhoilnL1PC1MPTooNRrGUttAWq90xsx/UZgUeTBIzX8xMcpV3r4t8DeiMihBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e23ba2e69dd4369ac12656d4947047414c6d69bd0cbcad7e02a6d00a76156307","last_reissued_at":"2026-07-05T04:15:21.410956Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:15:21.410956Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2204.08039","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":"2026-07-05T04:15:21Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KKNHmgzibEOOXNLithI/2Vxw3yHtDS5uQTc5RbtV3AQfXkQ8HfN7sLhHbz1YF/KDKVWZ4QKAdKYSj+SwY2ZYBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T00:42:42.785236Z"},"content_sha256":"71b0d2e59ef7ce746657076a044286f89d89248af0fa2fd2a90ba11940fab5ad","schema_version":"1.0","event_id":"sha256:71b0d2e59ef7ce746657076a044286f89d89248af0fa2fd2a90ba11940fab5ad"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:4I52FZU52Q3JVQJGK3KJI4CHIF","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Pathologies of Pre-trained Language Models in Few-shot Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ahmed Hassan Awadallah, Guoqing Zheng, Hanjie Chen, Yangfeng Ji","submitted_at":"2022-04-17T15:55:18Z","abstract_excerpt":"Although adapting pre-trained language models with few examples has shown promising performance on text classification, there is a lack of understanding of where the performance gain comes from. In this work, we propose to answer this question by interpreting the adaptation behavior using post-hoc explanations from model predictions. By modeling feature statistics of explanations, we discover that (1) without fine-tuning, pre-trained models (e.g. BERT and RoBERTa) show strong prediction bias across labels; (2) although few-shot fine-tuning can mitigate the prediction bias and demonstrate promi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.08039","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/2204.08039/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":"2026-07-05T04:15:21Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"H0m2PRilB2XPJKr6XtO1fc7h4G+qYAmxGZSDUW2HbzcLWS3+ge7LbeV2h52nrQAyCGeWAhMTxasrv4nDK9cwCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T00:42:42.785794Z"},"content_sha256":"2a416acc592243cd7aa83528cdd71955b1fc819f430668cc898345ae485eee35","schema_version":"1.0","event_id":"sha256:2a416acc592243cd7aa83528cdd71955b1fc819f430668cc898345ae485eee35"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4I52FZU52Q3JVQJGK3KJI4CHIF/bundle.json","state_url":"https://pith.science/pith/4I52FZU52Q3JVQJGK3KJI4CHIF/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4I52FZU52Q3JVQJGK3KJI4CHIF/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":"2026-08-19T00:42:42Z","links":{"resolver":"https://pith.science/pith/4I52FZU52Q3JVQJGK3KJI4CHIF","bundle":"https://pith.science/pith/4I52FZU52Q3JVQJGK3KJI4CHIF/bundle.json","state":"https://pith.science/pith/4I52FZU52Q3JVQJGK3KJI4CHIF/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4I52FZU52Q3JVQJGK3KJI4CHIF/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:4I52FZU52Q3JVQJGK3KJI4CHIF","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":"e7a20bccdca2b181bb8c23220607ee46b21a413e60657615fe2c9556cac11e76","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-17T15:55:18Z","title_canon_sha256":"1bdf236dc133ada1fba614dba8438fd6a9800a279f06fccca3d4b5ee61ee3812"},"schema_version":"1.0","source":{"id":"2204.08039","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2204.08039","created_at":"2026-07-05T04:15:21Z"},{"alias_kind":"arxiv_version","alias_value":"2204.08039v1","created_at":"2026-07-05T04:15:21Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.08039","created_at":"2026-07-05T04:15:21Z"},{"alias_kind":"pith_short_12","alias_value":"4I52FZU52Q3J","created_at":"2026-07-05T04:15:21Z"},{"alias_kind":"pith_short_16","alias_value":"4I52FZU52Q3JVQJG","created_at":"2026-07-05T04:15:21Z"},{"alias_kind":"pith_short_8","alias_value":"4I52FZU5","created_at":"2026-07-05T04:15:21Z"}],"graph_snapshots":[{"event_id":"sha256:2a416acc592243cd7aa83528cdd71955b1fc819f430668cc898345ae485eee35","target":"graph","created_at":"2026-07-05T04:15:21Z","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/2204.08039/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Although adapting pre-trained language models with few examples has shown promising performance on text classification, there is a lack of understanding of where the performance gain comes from. In this work, we propose to answer this question by interpreting the adaptation behavior using post-hoc explanations from model predictions. By modeling feature statistics of explanations, we discover that (1) without fine-tuning, pre-trained models (e.g. BERT and RoBERTa) show strong prediction bias across labels; (2) although few-shot fine-tuning can mitigate the prediction bias and demonstrate promi","authors_text":"Ahmed Hassan Awadallah, Guoqing Zheng, Hanjie Chen, Yangfeng Ji","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-17T15:55:18Z","title":"Pathologies of Pre-trained Language Models in Few-shot Fine-tuning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.08039","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":"sha256:71b0d2e59ef7ce746657076a044286f89d89248af0fa2fd2a90ba11940fab5ad","target":"record","created_at":"2026-07-05T04:15:21Z","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":"e7a20bccdca2b181bb8c23220607ee46b21a413e60657615fe2c9556cac11e76","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-17T15:55:18Z","title_canon_sha256":"1bdf236dc133ada1fba614dba8438fd6a9800a279f06fccca3d4b5ee61ee3812"},"schema_version":"1.0","source":{"id":"2204.08039","kind":"arxiv","version":1}},"canonical_sha256":"e23ba2e69dd4369ac12656d4947047414c6d69bd0cbcad7e02a6d00a76156307","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e23ba2e69dd4369ac12656d4947047414c6d69bd0cbcad7e02a6d00a76156307","first_computed_at":"2026-07-05T04:15:21.410956Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T04:15:21.410956Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"MzIvtiCzZnkuwgmeEaSjq0q9RFhoilnL1PC1MPTooNRrGUttAWq90xsx/UZgUeTBIzX8xMcpV3r4t8DeiMihBw==","signature_status":"signed_v1","signed_at":"2026-07-05T04:15:21.411297Z","signed_message":"canonical_sha256_bytes"},"source_id":"2204.08039","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:71b0d2e59ef7ce746657076a044286f89d89248af0fa2fd2a90ba11940fab5ad","sha256:2a416acc592243cd7aa83528cdd71955b1fc819f430668cc898345ae485eee35"],"state_sha256":"64e2089902affb12076d78f79339438a29c7fe901274fb94f5052970563f75f3"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"zu1LK/fSSk7MgKsNxJmH7P33qwgjlTd6u9KDkCFs5aeR0UCrEnpp/pMoKxrt5jbqBoV/MjJjsInGJmXyMJptDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-19T00:42:42.790946Z","bundle_sha256":"f4701a48688d58d0024f4e18506418aceee38fd3528a4fb92766cf5bbf285519"}}