{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:K72MCI25QXEHS4C5GWIK23MQL6","short_pith_number":"pith:K72MCI25","canonical_record":{"source":{"id":"2410.19278","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-25T03:21:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ee40973cd944ca113c266d461bf3a7b9b56ff13b6ad5c8725b04dd500f275a24","abstract_canon_sha256":"aafa03a7ff6f11faa2fb338fcb454f97e588f2528f0a10829fb015c31087c3c2"},"schema_version":"1.0"},"canonical_sha256":"57f4c1235d85c879705d3590ad6d905fb2ff6fb2ef04439824ec6ec28206e406","source":{"kind":"arxiv","id":"2410.19278","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.19278","created_at":"2026-07-05T09:30:23Z"},{"alias_kind":"arxiv_version","alias_value":"2410.19278v2","created_at":"2026-07-05T09:30:23Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.19278","created_at":"2026-07-05T09:30:23Z"},{"alias_kind":"pith_short_12","alias_value":"K72MCI25QXEH","created_at":"2026-07-05T09:30:23Z"},{"alias_kind":"pith_short_16","alias_value":"K72MCI25QXEHS4C5","created_at":"2026-07-05T09:30:23Z"},{"alias_kind":"pith_short_8","alias_value":"K72MCI25","created_at":"2026-07-05T09:30:23Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:K72MCI25QXEHS4C5GWIK23MQL6","target":"record","payload":{"canonical_record":{"source":{"id":"2410.19278","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-25T03:21:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ee40973cd944ca113c266d461bf3a7b9b56ff13b6ad5c8725b04dd500f275a24","abstract_canon_sha256":"aafa03a7ff6f11faa2fb338fcb454f97e588f2528f0a10829fb015c31087c3c2"},"schema_version":"1.0"},"canonical_sha256":"57f4c1235d85c879705d3590ad6d905fb2ff6fb2ef04439824ec6ec28206e406","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:30:23.085556Z","signature_b64":"7qscgeMlH7bG7Gsw7xMKfWYLmuSHm+zYzrQgFlgQro+dgvtyRD353fNv4MK73RKxQDG9LwWdMJkrb1vdLMqCDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"57f4c1235d85c879705d3590ad6d905fb2ff6fb2ef04439824ec6ec28206e406","last_reissued_at":"2026-07-05T09:30:23.085078Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:30:23.085078Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.19278","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:30:23Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rDxm08yYN2IXeiTPTKZwwYSeXKyIgWnDey/xMugLewP9MWOV5nOBOUByeHnZde9JUJkbCFrYAOytQBhhyAs1CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T06:38:30.790154Z"},"content_sha256":"3afe30c6791ff00cc98a0098df4af47323dc0b63a0173daceee3a42cd0c84d7d","schema_version":"1.0","event_id":"sha256:3afe30c6791ff00cc98a0098df4af47323dc0b63a0173daceee3a42cd0c84d7d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:K72MCI25QXEHS4C5GWIK23MQL6","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Applying sparse autoencoders to unlearn knowledge in language models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Arthur Conmy, Eoin Farrell, Yeu-Tong Lau","submitted_at":"2024-10-25T03:21:14Z","abstract_excerpt":"We investigate whether sparse autoencoders (SAEs) can be used to remove knowledge from language models. We use the biology subset of the Weapons of Mass Destruction Proxy dataset and test on the gemma-2b-it and gemma-2-2b-it language models. We demonstrate that individual interpretable biology-related SAE features can be used to unlearn a subset of WMDP-Bio questions with minimal side-effects in domains other than biology. Our results suggest that negative scaling of feature activations is necessary and that zero ablating features is ineffective. We find that intervening using multiple SAE fea"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.19278","kind":"arxiv","version":2},"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/2410.19278/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":"2026-07-05T09:30:23Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"OTjIOMiqBkeE20mraFOtc144U7FEjIWJgnKFGkq+QJktqxMFlMwZdYDOodBRQJLUvw2FtGhfbITpOeXlZprHBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T06:38:30.790653Z"},"content_sha256":"a05d3f1af1c381e7e71d37c7d2f2f0d3bfacc70c1076967aef5ea29c9fe38c4b","schema_version":"1.0","event_id":"sha256:a05d3f1af1c381e7e71d37c7d2f2f0d3bfacc70c1076967aef5ea29c9fe38c4b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/K72MCI25QXEHS4C5GWIK23MQL6/bundle.json","state_url":"https://pith.science/pith/K72MCI25QXEHS4C5GWIK23MQL6/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/K72MCI25QXEHS4C5GWIK23MQL6/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":"2026-08-09T06:38:30Z","links":{"resolver":"https://pith.science/pith/K72MCI25QXEHS4C5GWIK23MQL6","bundle":"https://pith.science/pith/K72MCI25QXEHS4C5GWIK23MQL6/bundle.json","state":"https://pith.science/pith/K72MCI25QXEHS4C5GWIK23MQL6/state.json","well_known_bundle":"https://pith.science/.well-known/pith/K72MCI25QXEHS4C5GWIK23MQL6/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:K72MCI25QXEHS4C5GWIK23MQL6","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":"aafa03a7ff6f11faa2fb338fcb454f97e588f2528f0a10829fb015c31087c3c2","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-25T03:21:14Z","title_canon_sha256":"ee40973cd944ca113c266d461bf3a7b9b56ff13b6ad5c8725b04dd500f275a24"},"schema_version":"1.0","source":{"id":"2410.19278","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.19278","created_at":"2026-07-05T09:30:23Z"},{"alias_kind":"arxiv_version","alias_value":"2410.19278v2","created_at":"2026-07-05T09:30:23Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.19278","created_at":"2026-07-05T09:30:23Z"},{"alias_kind":"pith_short_12","alias_value":"K72MCI25QXEH","created_at":"2026-07-05T09:30:23Z"},{"alias_kind":"pith_short_16","alias_value":"K72MCI25QXEHS4C5","created_at":"2026-07-05T09:30:23Z"},{"alias_kind":"pith_short_8","alias_value":"K72MCI25","created_at":"2026-07-05T09:30:23Z"}],"graph_snapshots":[{"event_id":"sha256:a05d3f1af1c381e7e71d37c7d2f2f0d3bfacc70c1076967aef5ea29c9fe38c4b","target":"graph","created_at":"2026-07-05T09:30:23Z","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/2410.19278/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We investigate whether sparse autoencoders (SAEs) can be used to remove knowledge from language models. We use the biology subset of the Weapons of Mass Destruction Proxy dataset and test on the gemma-2b-it and gemma-2-2b-it language models. We demonstrate that individual interpretable biology-related SAE features can be used to unlearn a subset of WMDP-Bio questions with minimal side-effects in domains other than biology. Our results suggest that negative scaling of feature activations is necessary and that zero ablating features is ineffective. We find that intervening using multiple SAE fea","authors_text":"Arthur Conmy, Eoin Farrell, Yeu-Tong Lau","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-25T03:21:14Z","title":"Applying sparse autoencoders to unlearn knowledge in language models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.19278","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":"sha256:3afe30c6791ff00cc98a0098df4af47323dc0b63a0173daceee3a42cd0c84d7d","target":"record","created_at":"2026-07-05T09:30:23Z","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":"aafa03a7ff6f11faa2fb338fcb454f97e588f2528f0a10829fb015c31087c3c2","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-25T03:21:14Z","title_canon_sha256":"ee40973cd944ca113c266d461bf3a7b9b56ff13b6ad5c8725b04dd500f275a24"},"schema_version":"1.0","source":{"id":"2410.19278","kind":"arxiv","version":2}},"canonical_sha256":"57f4c1235d85c879705d3590ad6d905fb2ff6fb2ef04439824ec6ec28206e406","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"57f4c1235d85c879705d3590ad6d905fb2ff6fb2ef04439824ec6ec28206e406","first_computed_at":"2026-07-05T09:30:23.085078Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:30:23.085078Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"7qscgeMlH7bG7Gsw7xMKfWYLmuSHm+zYzrQgFlgQro+dgvtyRD353fNv4MK73RKxQDG9LwWdMJkrb1vdLMqCDg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:30:23.085556Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.19278","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3afe30c6791ff00cc98a0098df4af47323dc0b63a0173daceee3a42cd0c84d7d","sha256:a05d3f1af1c381e7e71d37c7d2f2f0d3bfacc70c1076967aef5ea29c9fe38c4b"],"state_sha256":"5c2d875b99218f17221875d6d17a7454503c5411ca6278d024cd2de9ae290f1c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"hfpLX8vKyato4VWCs4W6pWloA6tpeom0O4M5hEGIsfIreL/wId/5SwYvwjlxDL4jHJnC3To7niZ8k3FaHmqSDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T06:38:30.795898Z","bundle_sha256":"8f6a86d6ef7cbcbcec39b8dbde16bf20f1f51614097fdc4024c8c8f16ca19fbe"}}