{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:ZBWLZWWXXZXIZPXSRFWPXRG3XW","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":"2dbc94be03d2cc4cf17f98ead21738454bdf564a8ecadee51b602e5230065a52","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-12-18T14:36:07Z","title_canon_sha256":"83849c295c42726e45ad7a47533b5741f1ba2caf07c69826ca1aaf852ea9a805"},"schema_version":"1.0","source":{"id":"2212.09095","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2212.09095","created_at":"2026-07-05T06:41:46Z"},{"alias_kind":"arxiv_version","alias_value":"2212.09095v2","created_at":"2026-07-05T06:41:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.09095","created_at":"2026-07-05T06:41:46Z"},{"alias_kind":"pith_short_12","alias_value":"ZBWLZWWXXZXI","created_at":"2026-07-05T06:41:46Z"},{"alias_kind":"pith_short_16","alias_value":"ZBWLZWWXXZXIZPXS","created_at":"2026-07-05T06:41:46Z"},{"alias_kind":"pith_short_8","alias_value":"ZBWLZWWX","created_at":"2026-07-05T06:41:46Z"}],"graph_snapshots":[{"event_id":"sha256:783123fdad6b205cc66581b0c9113cd332f6a6bb1afd05f50ee43024af50b105","target":"graph","created_at":"2026-07-05T06:41:46Z","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/2212.09095/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Language models have been shown to perform better with an increase in scale on a wide variety of tasks via the in-context learning paradigm. In this paper, we investigate the hypothesis that the ability of a large language model to in-context learn-perform a task is not uniformly spread across all of its underlying components. Using a 66 billion parameter language model (OPT-66B) across a diverse set of 14 downstream tasks, we find this is indeed the case: $\\sim$70% of attention heads and $\\sim$20% of feed forward networks can be removed with minimal decline in task performance. We find substa","authors_text":"Dan Roth, Hritik Bansal, Karthik Gopalakrishnan, Katrin Kirchhoff, Saket Dingliwal, Sravan Bodapati","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-12-18T14:36:07Z","title":"Rethinking the Role of Scale for In-Context Learning: An Interpretability-based Case Study at 66 Billion Scale"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.09095","kind":"arxiv","version":2},"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":"sha256:84f3471fdb210492b58bac6ac7a4af3d1a0ae645ee8307aea2134a244fa30ff8","target":"record","created_at":"2026-07-05T06:41:46Z","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":"2dbc94be03d2cc4cf17f98ead21738454bdf564a8ecadee51b602e5230065a52","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-12-18T14:36:07Z","title_canon_sha256":"83849c295c42726e45ad7a47533b5741f1ba2caf07c69826ca1aaf852ea9a805"},"schema_version":"1.0","source":{"id":"2212.09095","kind":"arxiv","version":2}},"canonical_sha256":"c86cbcdad7be6e8cbef2896cfbc4dbbd8815a365ba60836f9c944b6059a1970c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c86cbcdad7be6e8cbef2896cfbc4dbbd8815a365ba60836f9c944b6059a1970c","first_computed_at":"2026-07-05T06:41:46.571562Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T06:41:46.571562Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"0rpiX8KRJhivzrNO8mJW0nLpQ4zhlufzjB+f7H8hb5BB+CkqSXpBHaRxoAY8OkORJ6PWXkwNAK4k+sZ1ixSkBw==","signature_status":"signed_v1","signed_at":"2026-07-05T06:41:46.572060Z","signed_message":"canonical_sha256_bytes"},"source_id":"2212.09095","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:84f3471fdb210492b58bac6ac7a4af3d1a0ae645ee8307aea2134a244fa30ff8","sha256:783123fdad6b205cc66581b0c9113cd332f6a6bb1afd05f50ee43024af50b105"],"state_sha256":"24856aa084e5654570a450cf10b66731e103fada7d0f11fc249002ab80d61cc3"}