{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:WEDPYMIN5R6CLRMFBAB3SXGTLI","short_pith_number":"pith:WEDPYMIN","schema_version":"1.0","canonical_sha256":"b106fc310dec7c25c5850803b95cd35a20729d6abf6a2c07aa2e3ff2c1ef7adb","source":{"kind":"arxiv","id":"2004.03061","version":2},"attestation_state":"computed","paper":{"title":"Information-Theoretic Probing for Linguistic Structure","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Adina Williams, Josef Valvoda, Ran Zmigrod, Rowan Hall Maudslay, Ryan Cotterell, Tiago Pimentel","submitted_at":"2020-04-07T01:06:36Z","abstract_excerpt":"The success of neural networks on a diverse set of NLP tasks has led researchers to question how much these networks actually ``know'' about natural language. Probes are a natural way of assessing this. When probing, a researcher chooses a linguistic task and trains a supervised model to predict annotations in that linguistic task from the network's learned representations. If the probe does well, the researcher may conclude that the representations encode knowledge related to the task. A commonly held belief is that using simpler models as probes is better; the logic is that simpler models wi"},"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":"2004.03061","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-04-07T01:06:36Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"ff085fca879bc5d86349ea4ceb5fc0149bf88bd7221b93e98d7d72d4e71d83fa","abstract_canon_sha256":"13a5883306ccf5680f613ba1f6d3cecc81b76e22882bd9bc126f86fac50b8d6d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:05:12.308804Z","signature_b64":"bevOnq7HWWNS77XjIaE09AhIxWdVHR54n69alxNcxWcOicBn5nj2aQt+HZpyo6S/1MSjgSfqyQZTAByIl79LBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b106fc310dec7c25c5850803b95cd35a20729d6abf6a2c07aa2e3ff2c1ef7adb","last_reissued_at":"2026-07-05T01:05:12.308302Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:05:12.308302Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Information-Theoretic Probing for Linguistic Structure","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Adina Williams, Josef Valvoda, Ran Zmigrod, Rowan Hall Maudslay, Ryan Cotterell, Tiago Pimentel","submitted_at":"2020-04-07T01:06:36Z","abstract_excerpt":"The success of neural networks on a diverse set of NLP tasks has led researchers to question how much these networks actually ``know'' about natural language. Probes are a natural way of assessing this. When probing, a researcher chooses a linguistic task and trains a supervised model to predict annotations in that linguistic task from the network's learned representations. If the probe does well, the researcher may conclude that the representations encode knowledge related to the task. A commonly held belief is that using simpler models as probes is better; the logic is that simpler models wi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2004.03061","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/2004.03061/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":"2004.03061","created_at":"2026-07-05T01:05:12.308367+00:00"},{"alias_kind":"arxiv_version","alias_value":"2004.03061v2","created_at":"2026-07-05T01:05:12.308367+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2004.03061","created_at":"2026-07-05T01:05:12.308367+00:00"},{"alias_kind":"pith_short_12","alias_value":"WEDPYMIN5R6C","created_at":"2026-07-05T01:05:12.308367+00:00"},{"alias_kind":"pith_short_16","alias_value":"WEDPYMIN5R6CLRMF","created_at":"2026-07-05T01:05:12.308367+00:00"},{"alias_kind":"pith_short_8","alias_value":"WEDPYMIN","created_at":"2026-07-05T01:05:12.308367+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":199,"is_internal_anchor":false},{"citing_arxiv_id":"2606.11375","citing_title":"When Probing Accuracy Saturates, Fragility Resolves: A Complementary Metric for LLM Pre-Training Analysis","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":198,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07631","citing_title":"Inference Time Causal Probing in LLMs","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WEDPYMIN5R6CLRMFBAB3SXGTLI","json":"https://pith.science/pith/WEDPYMIN5R6CLRMFBAB3SXGTLI.json","graph_json":"https://pith.science/api/pith-number/WEDPYMIN5R6CLRMFBAB3SXGTLI/graph.json","events_json":"https://pith.science/api/pith-number/WEDPYMIN5R6CLRMFBAB3SXGTLI/events.json","paper":"https://pith.science/paper/WEDPYMIN"},"agent_actions":{"view_html":"https://pith.science/pith/WEDPYMIN5R6CLRMFBAB3SXGTLI","download_json":"https://pith.science/pith/WEDPYMIN5R6CLRMFBAB3SXGTLI.json","view_paper":"https://pith.science/paper/WEDPYMIN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2004.03061&json=true","fetch_graph":"https://pith.science/api/pith-number/WEDPYMIN5R6CLRMFBAB3SXGTLI/graph.json","fetch_events":"https://pith.science/api/pith-number/WEDPYMIN5R6CLRMFBAB3SXGTLI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WEDPYMIN5R6CLRMFBAB3SXGTLI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WEDPYMIN5R6CLRMFBAB3SXGTLI/action/storage_attestation","attest_author":"https://pith.science/pith/WEDPYMIN5R6CLRMFBAB3SXGTLI/action/author_attestation","sign_citation":"https://pith.science/pith/WEDPYMIN5R6CLRMFBAB3SXGTLI/action/citation_signature","submit_replication":"https://pith.science/pith/WEDPYMIN5R6CLRMFBAB3SXGTLI/action/replication_record"}},"created_at":"2026-07-05T01:05:12.308367+00:00","updated_at":"2026-07-05T01:05:12.308367+00:00"}