{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:BP6A5TMGZBYMIKACJLCD6LBRIV","short_pith_number":"pith:BP6A5TMG","canonical_record":{"source":{"id":"2501.09929","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-01-17T02:55:23Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"b0ed077b278e687b7328d8cf8d1a6c5d1ef7a56af98635de99b4b66a150a2608","abstract_canon_sha256":"ed9aefbe30c6786067baf0d3f6369a7d8b89a96592de8b754c543a8ce84290ee"},"schema_version":"1.0"},"canonical_sha256":"0bfc0ecd86c870c428024ac43f2c31456395334ef90af9d29500e9da2cbbbec0","source":{"kind":"arxiv","id":"2501.09929","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.09929","created_at":"2026-07-05T10:43:26Z"},{"alias_kind":"arxiv_version","alias_value":"2501.09929v3","created_at":"2026-07-05T10:43:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.09929","created_at":"2026-07-05T10:43:26Z"},{"alias_kind":"pith_short_12","alias_value":"BP6A5TMGZBYM","created_at":"2026-07-05T10:43:26Z"},{"alias_kind":"pith_short_16","alias_value":"BP6A5TMGZBYMIKAC","created_at":"2026-07-05T10:43:26Z"},{"alias_kind":"pith_short_8","alias_value":"BP6A5TMG","created_at":"2026-07-05T10:43:26Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:BP6A5TMGZBYMIKACJLCD6LBRIV","target":"record","payload":{"canonical_record":{"source":{"id":"2501.09929","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-01-17T02:55:23Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"b0ed077b278e687b7328d8cf8d1a6c5d1ef7a56af98635de99b4b66a150a2608","abstract_canon_sha256":"ed9aefbe30c6786067baf0d3f6369a7d8b89a96592de8b754c543a8ce84290ee"},"schema_version":"1.0"},"canonical_sha256":"0bfc0ecd86c870c428024ac43f2c31456395334ef90af9d29500e9da2cbbbec0","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:43:26.318052Z","signature_b64":"ayZKypl7nhB+JTv41oVhJbCLYs2tb2zb840yovWvAilwRegLQiOw1zjteJFbxnLU3tB5ZH0jWNAw6bhBOB5kBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0bfc0ecd86c870c428024ac43f2c31456395334ef90af9d29500e9da2cbbbec0","last_reissued_at":"2026-07-05T10:43:26.317575Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:43:26.317575Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2501.09929","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":"2026-07-05T10:43:26Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"JhEXvnERG21JLZr/AmmHR/lB9N5SnFtymauTvBydIPb5Hz51NlsNbKTvk5c1T95SWY6FpOS+jRD63k94XCNJBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-12T03:55:33.164102Z"},"content_sha256":"74dac58535684505137bdaae9a28272d05133b7843465b0a66cf0858526ee16b","schema_version":"1.0","event_id":"sha256:74dac58535684505137bdaae9a28272d05133b7843465b0a66cf0858526ee16b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:BP6A5TMGZBYMIKACJLCD6LBRIV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Interpretable Steering of Large Language Models with Feature Guided Activation Additions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Chandrasekaran Balaganesh, Chen Guang, Samuel Soo, Tan Guoxian, Wesley Teng, Yan Ming","submitted_at":"2025-01-17T02:55:23Z","abstract_excerpt":"Effective and reliable control over large language model (LLM) behavior is a significant challenge. While activation steering methods, which add steering vectors to a model's hidden states, are a promising approach, existing techniques often lack precision and interpretability in how they influence model outputs. We introduce Feature Guided Activation Additions (FGAA), a novel activation steering method that leverages insights from Contrastive Activation Addition (CAA) and Sparse Autoencoder-Targeted Steering (SAE-TS). By operating in the latent space of a Sparse Autoencoder (SAE) and employin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.09929","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/2501.09929/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":"2026-07-05T10:43:26Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Pc7oWZYscMflM25LVtdZaKS9umVluKyato+bsM7txGZTj1JZIY+o/z2XdAGq/gQTCV1K+BhC/D0HYdHIy8giDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-12T03:55:33.164486Z"},"content_sha256":"2178805852b24f5f72efb4d6e4730dbdf835cc9f8395101da7d24c04380d5850","schema_version":"1.0","event_id":"sha256:2178805852b24f5f72efb4d6e4730dbdf835cc9f8395101da7d24c04380d5850"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BP6A5TMGZBYMIKACJLCD6LBRIV/bundle.json","state_url":"https://pith.science/pith/BP6A5TMGZBYMIKACJLCD6LBRIV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BP6A5TMGZBYMIKACJLCD6LBRIV/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":"2026-07-12T03:55:33Z","links":{"resolver":"https://pith.science/pith/BP6A5TMGZBYMIKACJLCD6LBRIV","bundle":"https://pith.science/pith/BP6A5TMGZBYMIKACJLCD6LBRIV/bundle.json","state":"https://pith.science/pith/BP6A5TMGZBYMIKACJLCD6LBRIV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BP6A5TMGZBYMIKACJLCD6LBRIV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:BP6A5TMGZBYMIKACJLCD6LBRIV","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":"ed9aefbe30c6786067baf0d3f6369a7d8b89a96592de8b754c543a8ce84290ee","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-01-17T02:55:23Z","title_canon_sha256":"b0ed077b278e687b7328d8cf8d1a6c5d1ef7a56af98635de99b4b66a150a2608"},"schema_version":"1.0","source":{"id":"2501.09929","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.09929","created_at":"2026-07-05T10:43:26Z"},{"alias_kind":"arxiv_version","alias_value":"2501.09929v3","created_at":"2026-07-05T10:43:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.09929","created_at":"2026-07-05T10:43:26Z"},{"alias_kind":"pith_short_12","alias_value":"BP6A5TMGZBYM","created_at":"2026-07-05T10:43:26Z"},{"alias_kind":"pith_short_16","alias_value":"BP6A5TMGZBYMIKAC","created_at":"2026-07-05T10:43:26Z"},{"alias_kind":"pith_short_8","alias_value":"BP6A5TMG","created_at":"2026-07-05T10:43:26Z"}],"graph_snapshots":[{"event_id":"sha256:2178805852b24f5f72efb4d6e4730dbdf835cc9f8395101da7d24c04380d5850","target":"graph","created_at":"2026-07-05T10:43:26Z","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/2501.09929/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Effective and reliable control over large language model (LLM) behavior is a significant challenge. While activation steering methods, which add steering vectors to a model's hidden states, are a promising approach, existing techniques often lack precision and interpretability in how they influence model outputs. We introduce Feature Guided Activation Additions (FGAA), a novel activation steering method that leverages insights from Contrastive Activation Addition (CAA) and Sparse Autoencoder-Targeted Steering (SAE-TS). By operating in the latent space of a Sparse Autoencoder (SAE) and employin","authors_text":"Chandrasekaran Balaganesh, Chen Guang, Samuel Soo, Tan Guoxian, Wesley Teng, Yan Ming","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-01-17T02:55:23Z","title":"Interpretable Steering of Large Language Models with Feature Guided Activation Additions"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.09929","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":"sha256:74dac58535684505137bdaae9a28272d05133b7843465b0a66cf0858526ee16b","target":"record","created_at":"2026-07-05T10:43:26Z","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":"ed9aefbe30c6786067baf0d3f6369a7d8b89a96592de8b754c543a8ce84290ee","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-01-17T02:55:23Z","title_canon_sha256":"b0ed077b278e687b7328d8cf8d1a6c5d1ef7a56af98635de99b4b66a150a2608"},"schema_version":"1.0","source":{"id":"2501.09929","kind":"arxiv","version":3}},"canonical_sha256":"0bfc0ecd86c870c428024ac43f2c31456395334ef90af9d29500e9da2cbbbec0","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0bfc0ecd86c870c428024ac43f2c31456395334ef90af9d29500e9da2cbbbec0","first_computed_at":"2026-07-05T10:43:26.317575Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:43:26.317575Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ayZKypl7nhB+JTv41oVhJbCLYs2tb2zb840yovWvAilwRegLQiOw1zjteJFbxnLU3tB5ZH0jWNAw6bhBOB5kBw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:43:26.318052Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.09929","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:74dac58535684505137bdaae9a28272d05133b7843465b0a66cf0858526ee16b","sha256:2178805852b24f5f72efb4d6e4730dbdf835cc9f8395101da7d24c04380d5850"],"state_sha256":"fdd5936b3380f70b9deb6cdf73ff5d4cd132a36ee8f95afc5444650d21a3d9ed"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1ptqkMK3HrSqjAEUBMAV8zt1A+O4+NxhbJozPY7CEQDS5EuVN/aqEZT+lWBc/T3vU/909L2J6jH5uHDaYlrZCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-12T03:55:33.166634Z","bundle_sha256":"0c87e85ff57e96909e92f201d663d2ea946e31c7de77478799b63c7dba932913"}}