{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:GV7UZRHYGSLSBELXZRO6I2CLX7","short_pith_number":"pith:GV7UZRHY","canonical_record":{"source":{"id":"2501.17148","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-28T18:51:24Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"c768c5e61731523f55ff727450b0c2a56bbbecc65f10d8976d8910555c707d90","abstract_canon_sha256":"a02a5d3ad41c49e3131ffadd79da242f2241a923989838a9c76912b5d74c6255"},"schema_version":"1.0"},"canonical_sha256":"357f4cc4f83497209177cc5de4684bbfe255303ae0c1a474ce23b13688b48390","source":{"kind":"arxiv","id":"2501.17148","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.17148","created_at":"2026-07-05T10:23:28Z"},{"alias_kind":"arxiv_version","alias_value":"2501.17148v3","created_at":"2026-07-05T10:23:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.17148","created_at":"2026-07-05T10:23:28Z"},{"alias_kind":"pith_short_12","alias_value":"GV7UZRHYGSLS","created_at":"2026-07-05T10:23:28Z"},{"alias_kind":"pith_short_16","alias_value":"GV7UZRHYGSLSBELX","created_at":"2026-07-05T10:23:28Z"},{"alias_kind":"pith_short_8","alias_value":"GV7UZRHY","created_at":"2026-07-05T10:23:28Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:GV7UZRHYGSLSBELXZRO6I2CLX7","target":"record","payload":{"canonical_record":{"source":{"id":"2501.17148","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-28T18:51:24Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"c768c5e61731523f55ff727450b0c2a56bbbecc65f10d8976d8910555c707d90","abstract_canon_sha256":"a02a5d3ad41c49e3131ffadd79da242f2241a923989838a9c76912b5d74c6255"},"schema_version":"1.0"},"canonical_sha256":"357f4cc4f83497209177cc5de4684bbfe255303ae0c1a474ce23b13688b48390","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:23:28.862350Z","signature_b64":"Ca4Bmnj1JNQO09ELXJ8dhMb4ssAWu9uO4oMDpSGqjj2XbllpyxMuYkm037rXr/zMTlmbxE3tbVU+43asOF9pCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"357f4cc4f83497209177cc5de4684bbfe255303ae0c1a474ce23b13688b48390","last_reissued_at":"2026-07-05T10:23:28.861853Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:23:28.861853Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2501.17148","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:23:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"AYRF1g3v3qrtdzIQupB7HxP66WbOcxgvgmOst87GvaDfr9q6GEbwKqXLj63aGpMsXsFL/TRCIOyW2lbqUZSMDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T11:39:28.327125Z"},"content_sha256":"b3803a35cc3e5d863823f7be9e0d1afc6744df60e10633978ef0c63c818b6287","schema_version":"1.0","event_id":"sha256:b3803a35cc3e5d863823f7be9e0d1afc6744df60e10633978ef0c63c818b6287"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:GV7UZRHYGSLSBELXZRO6I2CLX7","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aryaman Arora, Atticus Geiger, Christopher D. Manning, Christopher Potts, Dan Jurafsky, Jing Huang, Zheng Wang, Zhengxuan Wu","submitted_at":"2025-01-28T18:51:24Z","abstract_excerpt":"Fine-grained steering of language model outputs is essential for safety and reliability. Prompting and finetuning are widely used to achieve these goals, but interpretability researchers have proposed a variety of representation-based techniques as well, including sparse autoencoders (SAEs), linear artificial tomography, supervised steering vectors, linear probes, and representation finetuning. At present, there is no benchmark for making direct comparisons between these proposals. Therefore, we introduce AxBench, a large-scale benchmark for steering and concept detection, and report experimen"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.17148","kind":"arxiv","version":3},"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/2501.17148/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":"2026-07-05T10:23:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"WrTJLZcTydx/fs7etjFm+8FvIiqFCfTI+wbulpEfN7lbMtJFqLYqQHsbYUrt+dDYFz+NP3cfTHGdIT2fJmtTDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T11:39:28.327600Z"},"content_sha256":"4c6ae67e9b05bd14c1048d1d3d726785932995e6e881e062a33258e30b1bd383","schema_version":"1.0","event_id":"sha256:4c6ae67e9b05bd14c1048d1d3d726785932995e6e881e062a33258e30b1bd383"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/GV7UZRHYGSLSBELXZRO6I2CLX7/bundle.json","state_url":"https://pith.science/pith/GV7UZRHYGSLSBELXZRO6I2CLX7/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/GV7UZRHYGSLSBELXZRO6I2CLX7/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":"2026-08-07T11:39:28Z","links":{"resolver":"https://pith.science/pith/GV7UZRHYGSLSBELXZRO6I2CLX7","bundle":"https://pith.science/pith/GV7UZRHYGSLSBELXZRO6I2CLX7/bundle.json","state":"https://pith.science/pith/GV7UZRHYGSLSBELXZRO6I2CLX7/state.json","well_known_bundle":"https://pith.science/.well-known/pith/GV7UZRHYGSLSBELXZRO6I2CLX7/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:GV7UZRHYGSLSBELXZRO6I2CLX7","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":"a02a5d3ad41c49e3131ffadd79da242f2241a923989838a9c76912b5d74c6255","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-28T18:51:24Z","title_canon_sha256":"c768c5e61731523f55ff727450b0c2a56bbbecc65f10d8976d8910555c707d90"},"schema_version":"1.0","source":{"id":"2501.17148","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.17148","created_at":"2026-07-05T10:23:28Z"},{"alias_kind":"arxiv_version","alias_value":"2501.17148v3","created_at":"2026-07-05T10:23:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.17148","created_at":"2026-07-05T10:23:28Z"},{"alias_kind":"pith_short_12","alias_value":"GV7UZRHYGSLS","created_at":"2026-07-05T10:23:28Z"},{"alias_kind":"pith_short_16","alias_value":"GV7UZRHYGSLSBELX","created_at":"2026-07-05T10:23:28Z"},{"alias_kind":"pith_short_8","alias_value":"GV7UZRHY","created_at":"2026-07-05T10:23:28Z"}],"graph_snapshots":[{"event_id":"sha256:4c6ae67e9b05bd14c1048d1d3d726785932995e6e881e062a33258e30b1bd383","target":"graph","created_at":"2026-07-05T10:23:28Z","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/2501.17148/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Fine-grained steering of language model outputs is essential for safety and reliability. Prompting and finetuning are widely used to achieve these goals, but interpretability researchers have proposed a variety of representation-based techniques as well, including sparse autoencoders (SAEs), linear artificial tomography, supervised steering vectors, linear probes, and representation finetuning. At present, there is no benchmark for making direct comparisons between these proposals. Therefore, we introduce AxBench, a large-scale benchmark for steering and concept detection, and report experimen","authors_text":"Aryaman Arora, Atticus Geiger, Christopher D. Manning, Christopher Potts, Dan Jurafsky, Jing Huang, Zheng Wang, Zhengxuan Wu","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-28T18:51:24Z","title":"AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.17148","kind":"arxiv","version":3},"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:b3803a35cc3e5d863823f7be9e0d1afc6744df60e10633978ef0c63c818b6287","target":"record","created_at":"2026-07-05T10:23:28Z","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":"a02a5d3ad41c49e3131ffadd79da242f2241a923989838a9c76912b5d74c6255","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-28T18:51:24Z","title_canon_sha256":"c768c5e61731523f55ff727450b0c2a56bbbecc65f10d8976d8910555c707d90"},"schema_version":"1.0","source":{"id":"2501.17148","kind":"arxiv","version":3}},"canonical_sha256":"357f4cc4f83497209177cc5de4684bbfe255303ae0c1a474ce23b13688b48390","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"357f4cc4f83497209177cc5de4684bbfe255303ae0c1a474ce23b13688b48390","first_computed_at":"2026-07-05T10:23:28.861853Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:23:28.861853Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Ca4Bmnj1JNQO09ELXJ8dhMb4ssAWu9uO4oMDpSGqjj2XbllpyxMuYkm037rXr/zMTlmbxE3tbVU+43asOF9pCw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:23:28.862350Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.17148","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b3803a35cc3e5d863823f7be9e0d1afc6744df60e10633978ef0c63c818b6287","sha256:4c6ae67e9b05bd14c1048d1d3d726785932995e6e881e062a33258e30b1bd383"],"state_sha256":"23758d1a1b17ff7e6814c80f6acca20d48500da5fabc92f795c5b5b193da3758"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"UnIhufMYa0nSiAlqetC0jsXvsMTIhO5xNUSoo2elXQxDweVWGPCRoL8kmRzzgv+niWFIJfN+FxzuIW0+Zfq1Dg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T11:39:28.331569Z","bundle_sha256":"26e3686b5d2218948470edca4b2bac254cac70c0f3624b32b3e90ae087b57f83"}}