{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:FC764B54GGEZM74TCKTOOFKLGX","short_pith_number":"pith:FC764B54","schema_version":"1.0","canonical_sha256":"28bfee07bc3189967f9312a6e7154b35ce7a8468ed764600938eafde8008c972","source":{"kind":"arxiv","id":"2106.00085","version":3},"attestation_state":"computed","paper":{"title":"Language Model Evaluation Beyond Perplexity","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Clara Meister, Ryan Cotterell","submitted_at":"2021-05-31T20:13:44Z","abstract_excerpt":"We propose an alternate approach to quantifying how well language models learn natural language: we ask how well they match the statistical tendencies of natural language. To answer this question, we analyze whether text generated from language models exhibits the statistical tendencies present in the human-generated text on which they were trained. We provide a framework--paired with significance tests--for evaluating the fit of language models to these trends. We find that neural language models appear to learn only a subset of the tendencies considered, but align much more closely with empi"},"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":"2106.00085","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-05-31T20:13:44Z","cross_cats_sorted":[],"title_canon_sha256":"16f3f1b3f5919c0debcfd8ce9697c22bdeae22826967ab70c333d4027ccbbd58","abstract_canon_sha256":"746ad0758b0463fc314381a87f54ae98c474774ba243a696c15fb2a894c9dee5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:09:46.118143Z","signature_b64":"k6NtWQhBLnIMJu1IHzkwWsVYhG3hiJmwyD4eLAjFoIZOcC253s3sGxH2x5LWcFOvUmDcjvUlJWlg5Xs639qWDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"28bfee07bc3189967f9312a6e7154b35ce7a8468ed764600938eafde8008c972","last_reissued_at":"2026-07-05T03:09:46.117659Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:09:46.117659Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Language Model Evaluation Beyond Perplexity","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Clara Meister, Ryan Cotterell","submitted_at":"2021-05-31T20:13:44Z","abstract_excerpt":"We propose an alternate approach to quantifying how well language models learn natural language: we ask how well they match the statistical tendencies of natural language. To answer this question, we analyze whether text generated from language models exhibits the statistical tendencies present in the human-generated text on which they were trained. We provide a framework--paired with significance tests--for evaluating the fit of language models to these trends. We find that neural language models appear to learn only a subset of the tendencies considered, but align much more closely with empi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.00085","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/2106.00085/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":"2106.00085","created_at":"2026-07-05T03:09:46.117722+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.00085v3","created_at":"2026-07-05T03:09:46.117722+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.00085","created_at":"2026-07-05T03:09:46.117722+00:00"},{"alias_kind":"pith_short_12","alias_value":"FC764B54GGEZ","created_at":"2026-07-05T03:09:46.117722+00:00"},{"alias_kind":"pith_short_16","alias_value":"FC764B54GGEZM74T","created_at":"2026-07-05T03:09:46.117722+00:00"},{"alias_kind":"pith_short_8","alias_value":"FC764B54","created_at":"2026-07-05T03:09:46.117722+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":189,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":188,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10390","citing_title":"LLM-PRISM: Characterizing Silent Data Corruption from Permanent GPU Faults in LLM Training","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FC764B54GGEZM74TCKTOOFKLGX","json":"https://pith.science/pith/FC764B54GGEZM74TCKTOOFKLGX.json","graph_json":"https://pith.science/api/pith-number/FC764B54GGEZM74TCKTOOFKLGX/graph.json","events_json":"https://pith.science/api/pith-number/FC764B54GGEZM74TCKTOOFKLGX/events.json","paper":"https://pith.science/paper/FC764B54"},"agent_actions":{"view_html":"https://pith.science/pith/FC764B54GGEZM74TCKTOOFKLGX","download_json":"https://pith.science/pith/FC764B54GGEZM74TCKTOOFKLGX.json","view_paper":"https://pith.science/paper/FC764B54","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.00085&json=true","fetch_graph":"https://pith.science/api/pith-number/FC764B54GGEZM74TCKTOOFKLGX/graph.json","fetch_events":"https://pith.science/api/pith-number/FC764B54GGEZM74TCKTOOFKLGX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FC764B54GGEZM74TCKTOOFKLGX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FC764B54GGEZM74TCKTOOFKLGX/action/storage_attestation","attest_author":"https://pith.science/pith/FC764B54GGEZM74TCKTOOFKLGX/action/author_attestation","sign_citation":"https://pith.science/pith/FC764B54GGEZM74TCKTOOFKLGX/action/citation_signature","submit_replication":"https://pith.science/pith/FC764B54GGEZM74TCKTOOFKLGX/action/replication_record"}},"created_at":"2026-07-05T03:09:46.117722+00:00","updated_at":"2026-07-05T03:09:46.117722+00:00"}