{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ZEE6QVIPVD6L7QNPYOEZWQ2BMG","short_pith_number":"pith:ZEE6QVIP","schema_version":"1.0","canonical_sha256":"c909e8550fa8fcbfc1afc3899b434161bc747e100c938a555adf9500b5ff5a97","source":{"kind":"arxiv","id":"2404.12957","version":2},"attestation_state":"computed","paper":{"title":"Towards Reliable Latent Knowledge Estimation in LLMs: Zero-Prompt Many-Shot Based Factual Knowledge Extraction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Bishwamittra Ghosh, Camila Kolling, Evimaria Terzi, Krishna P. Gummadi, Laurent Bindschaedler, Mohammad Aflah Khan, Qinyuan Wu, Soumi Das, Till Speicher, Vedant Nanda","submitted_at":"2024-04-19T15:40:39Z","abstract_excerpt":"In this paper, we focus on the challenging task of reliably estimating factual knowledge that is embedded inside large language models (LLMs). To avoid reliability concerns with prior approaches, we propose to eliminate prompt engineering when probing LLMs for factual knowledge. Our approach, called Zero-Prompt Latent Knowledge Estimator (ZP-LKE), leverages the in-context learning ability of LLMs to communicate both the factual knowledge question as well as the expected answer format. Our knowledge estimator is both conceptually simpler (i.e., doesn't depend on meta-linguistic judgments of LLM"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2404.12957","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-19T15:40:39Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"fc39bb4327b0f447f47a22e3a768c0e47395e19de4ad2aee974f05d73c677f51","abstract_canon_sha256":"31a113390e02aba069d2fbb67f270953ab65fd9212c92326087539ce499e6db5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:50:31.029378Z","signature_b64":"cDeuKT8S9CmQ7m8JoTYMTuLuVNQzIGMb3Smr6Q+BsAx+z3VUZ5GNiBwbwc5UoIXnUHZOvlj+kNRxSmcMmnjwCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c909e8550fa8fcbfc1afc3899b434161bc747e100c938a555adf9500b5ff5a97","last_reissued_at":"2026-07-05T09:50:31.028897Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:50:31.028897Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Reliable Latent Knowledge Estimation in LLMs: Zero-Prompt Many-Shot Based Factual Knowledge Extraction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Bishwamittra Ghosh, Camila Kolling, Evimaria Terzi, Krishna P. Gummadi, Laurent Bindschaedler, Mohammad Aflah Khan, Qinyuan Wu, Soumi Das, Till Speicher, Vedant Nanda","submitted_at":"2024-04-19T15:40:39Z","abstract_excerpt":"In this paper, we focus on the challenging task of reliably estimating factual knowledge that is embedded inside large language models (LLMs). To avoid reliability concerns with prior approaches, we propose to eliminate prompt engineering when probing LLMs for factual knowledge. Our approach, called Zero-Prompt Latent Knowledge Estimator (ZP-LKE), leverages the in-context learning ability of LLMs to communicate both the factual knowledge question as well as the expected answer format. Our knowledge estimator is both conceptually simpler (i.e., doesn't depend on meta-linguistic judgments of LLM"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.12957","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/2404.12957/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2404.12957","created_at":"2026-07-05T09:50:31.028951+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.12957v2","created_at":"2026-07-05T09:50:31.028951+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.12957","created_at":"2026-07-05T09:50:31.028951+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZEE6QVIPVD6L","created_at":"2026-07-05T09:50:31.028951+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZEE6QVIPVD6L7QNP","created_at":"2026-07-05T09:50:31.028951+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZEE6QVIP","created_at":"2026-07-05T09:50:31.028951+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.19137","citing_title":"Construction of Knowledge Graph based on Language Model","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZEE6QVIPVD6L7QNPYOEZWQ2BMG","json":"https://pith.science/pith/ZEE6QVIPVD6L7QNPYOEZWQ2BMG.json","graph_json":"https://pith.science/api/pith-number/ZEE6QVIPVD6L7QNPYOEZWQ2BMG/graph.json","events_json":"https://pith.science/api/pith-number/ZEE6QVIPVD6L7QNPYOEZWQ2BMG/events.json","paper":"https://pith.science/paper/ZEE6QVIP"},"agent_actions":{"view_html":"https://pith.science/pith/ZEE6QVIPVD6L7QNPYOEZWQ2BMG","download_json":"https://pith.science/pith/ZEE6QVIPVD6L7QNPYOEZWQ2BMG.json","view_paper":"https://pith.science/paper/ZEE6QVIP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.12957&json=true","fetch_graph":"https://pith.science/api/pith-number/ZEE6QVIPVD6L7QNPYOEZWQ2BMG/graph.json","fetch_events":"https://pith.science/api/pith-number/ZEE6QVIPVD6L7QNPYOEZWQ2BMG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZEE6QVIPVD6L7QNPYOEZWQ2BMG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZEE6QVIPVD6L7QNPYOEZWQ2BMG/action/storage_attestation","attest_author":"https://pith.science/pith/ZEE6QVIPVD6L7QNPYOEZWQ2BMG/action/author_attestation","sign_citation":"https://pith.science/pith/ZEE6QVIPVD6L7QNPYOEZWQ2BMG/action/citation_signature","submit_replication":"https://pith.science/pith/ZEE6QVIPVD6L7QNPYOEZWQ2BMG/action/replication_record"}},"created_at":"2026-07-05T09:50:31.028951+00:00","updated_at":"2026-07-05T09:50:31.028951+00:00"}