{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:DQDI5YG7HPPFD7XUVZNG35FHLW","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":"166fccc59241f05f70483e5431e0a7a10c63a37c39ffa7cd88b3eb12766618f9","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-21T02:45:11Z","title_canon_sha256":"9d92e654db5178568cdfaaba090d63ecc7a400cc7ec0160ef0630bb1156aea5a"},"schema_version":"1.0","source":{"id":"2505.15038","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.15038","created_at":"2026-07-05T11:45:14Z"},{"alias_kind":"arxiv_version","alias_value":"2505.15038v2","created_at":"2026-07-05T11:45:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.15038","created_at":"2026-07-05T11:45:14Z"},{"alias_kind":"pith_short_12","alias_value":"DQDI5YG7HPPF","created_at":"2026-07-05T11:45:14Z"},{"alias_kind":"pith_short_16","alias_value":"DQDI5YG7HPPFD7XU","created_at":"2026-07-05T11:45:14Z"},{"alias_kind":"pith_short_8","alias_value":"DQDI5YG7","created_at":"2026-07-05T11:45:14Z"}],"graph_snapshots":[{"event_id":"sha256:5d7107ac817b776739b51d4d86728244aa3b3b0db20e6d0ab9761e15b52f94e1","target":"graph","created_at":"2026-07-05T11:45:14Z","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/2505.15038/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Linear concept vectors effectively steer LLMs, but existing methods suffer from noisy features in diverse datasets that undermine steering robustness. We propose Sparse Autoencoder-Denoised Concept Vectors (SDCV), which selectively keep the most discriminative SAE latents while reconstructing hidden representations. Our key insight is that concept-relevant signals can be explicitly separated from dataset noise by scaling up activations of top-k latents that best differentiate positive and negative samples. Applied to linear probing and difference-in-mean, SDCV consistently improves steering su","authors_text":"Bo Shen, Fan Yang, Haiyan Zhao, Mengnan Du, Ninghao Liu, Xuansheng Wu","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-21T02:45:11Z","title":"Denoising Concept Vectors with Sparse Autoencoders for Improved Language Model Steering"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.15038","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:efc72ee8010ac87ef62565a61da000b27fda6284b7b4261cc1f1bf3eb3e42c51","target":"record","created_at":"2026-07-05T11:45:14Z","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":"166fccc59241f05f70483e5431e0a7a10c63a37c39ffa7cd88b3eb12766618f9","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-21T02:45:11Z","title_canon_sha256":"9d92e654db5178568cdfaaba090d63ecc7a400cc7ec0160ef0630bb1156aea5a"},"schema_version":"1.0","source":{"id":"2505.15038","kind":"arxiv","version":2}},"canonical_sha256":"1c068ee0df3bde51fef4ae5a6df4a75d97b7d8e19e377cbb5c351c37aa8ad19b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1c068ee0df3bde51fef4ae5a6df4a75d97b7d8e19e377cbb5c351c37aa8ad19b","first_computed_at":"2026-07-05T11:45:14.929854Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:45:14.929854Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"NTfvPR/z18SZyTJx6cXnSh4Ikk2NVywon+OAwJgqhkGp+mCu9x0Z2ivJCwF7oH9rx3LP2gdnAjnCFDQ6HbiVAg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:45:14.930392Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.15038","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:efc72ee8010ac87ef62565a61da000b27fda6284b7b4261cc1f1bf3eb3e42c51","sha256:5d7107ac817b776739b51d4d86728244aa3b3b0db20e6d0ab9761e15b52f94e1"],"state_sha256":"549ca7275ade553370abdf5786adf6827796056e291f4d2b0e753616a085ef17"}