{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2020:OAD2UOTP2M4N7MM7MQHBGMABC3","short_pith_number":"pith:OAD2UOTP","canonical_record":{"source":{"id":"2001.08764","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-01-23T19:06:18Z","cross_cats_sorted":[],"title_canon_sha256":"2a728d844524ce67f9839db9b1e3e11de0164f8e73a6ee08a6e434b4f2421df2","abstract_canon_sha256":"319dd14409b45055ae6ce2bf6b110cde5118fd2a7a9a2b32d8030828ad213ba3"},"schema_version":"1.0"},"canonical_sha256":"7007aa3a6fd338dfb19f640e13300116c37e9a09437d78a17692f82606d1c4d5","source":{"kind":"arxiv","id":"2001.08764","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2001.08764","created_at":"2026-07-05T01:47:42Z"},{"alias_kind":"arxiv_version","alias_value":"2001.08764v2","created_at":"2026-07-05T01:47:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2001.08764","created_at":"2026-07-05T01:47:42Z"},{"alias_kind":"pith_short_12","alias_value":"OAD2UOTP2M4N","created_at":"2026-07-05T01:47:42Z"},{"alias_kind":"pith_short_16","alias_value":"OAD2UOTP2M4N7MM7","created_at":"2026-07-05T01:47:42Z"},{"alias_kind":"pith_short_8","alias_value":"OAD2UOTP","created_at":"2026-07-05T01:47:42Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2020:OAD2UOTP2M4N7MM7MQHBGMABC3","target":"record","payload":{"canonical_record":{"source":{"id":"2001.08764","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-01-23T19:06:18Z","cross_cats_sorted":[],"title_canon_sha256":"2a728d844524ce67f9839db9b1e3e11de0164f8e73a6ee08a6e434b4f2421df2","abstract_canon_sha256":"319dd14409b45055ae6ce2bf6b110cde5118fd2a7a9a2b32d8030828ad213ba3"},"schema_version":"1.0"},"canonical_sha256":"7007aa3a6fd338dfb19f640e13300116c37e9a09437d78a17692f82606d1c4d5","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:47:42.824522Z","signature_b64":"/D6935ZM/Zi1cLWbgQ+ALNxr+640nfNGY2fwlluof4gklo3u2mKjnN5aD45PmEIi3YJhsQM8w+u2hIrk0+6SCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7007aa3a6fd338dfb19f640e13300116c37e9a09437d78a17692f82606d1c4d5","last_reissued_at":"2026-07-05T01:47:42.824121Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:47:42.824121Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2001.08764","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T01:47:42Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NZPnSEsAXlZNTKhk/GQxRsAIBvK0UvmlGU2NM6FGQAhf806aA8sHf/goEox4JE4ypwdvQcespdZNQh9fgU+WBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T10:20:10.846572Z"},"content_sha256":"ddd1faba095b928cc211a15428a2e96f6dd87e4868309983f8cb37f2e323f9d0","schema_version":"1.0","event_id":"sha256:ddd1faba095b928cc211a15428a2e96f6dd87e4868309983f8cb37f2e323f9d0"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2020:OAD2UOTP2M4N7MM7MQHBGMABC3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Reducing Non-Normative Text Generation from Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Mark Riedl, Siyan Li, Spencer Frazier, Xiangyu Peng","submitted_at":"2020-01-23T19:06:18Z","abstract_excerpt":"Large-scale, transformer-based language models such as GPT-2 are pretrained on diverse corpora scraped from the internet. Consequently, they are prone to generating non-normative text (i.e. in violation of social norms). We introduce a technique for fine-tuning GPT-2, using a policy gradient reinforcement learning technique and a normative text classifier to produce reward and punishment values. We evaluate our technique on five data sets using automated and human participant experiments. The normative text classifier is 81-90% accurate when compared to gold-standard human judgments of normati"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2001.08764","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/2001.08764/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":"2026-07-05T01:47:42Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rACHLQPqgYbqMLrb6O6hpswzoHM2xPqwkTiCzrJ8VarPGhigxvfXkgHRiUElOfBLKoYC1U9cqtZpCQ9iI9IVDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T10:20:10.846943Z"},"content_sha256":"c6265490d23db770ec0047de77bf17a90540a02e4bd24a39daec218f2ddece95","schema_version":"1.0","event_id":"sha256:c6265490d23db770ec0047de77bf17a90540a02e4bd24a39daec218f2ddece95"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/OAD2UOTP2M4N7MM7MQHBGMABC3/bundle.json","state_url":"https://pith.science/pith/OAD2UOTP2M4N7MM7MQHBGMABC3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/OAD2UOTP2M4N7MM7MQHBGMABC3/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":"2026-07-07T10:20:10Z","links":{"resolver":"https://pith.science/pith/OAD2UOTP2M4N7MM7MQHBGMABC3","bundle":"https://pith.science/pith/OAD2UOTP2M4N7MM7MQHBGMABC3/bundle.json","state":"https://pith.science/pith/OAD2UOTP2M4N7MM7MQHBGMABC3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/OAD2UOTP2M4N7MM7MQHBGMABC3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2020:OAD2UOTP2M4N7MM7MQHBGMABC3","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":"319dd14409b45055ae6ce2bf6b110cde5118fd2a7a9a2b32d8030828ad213ba3","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-01-23T19:06:18Z","title_canon_sha256":"2a728d844524ce67f9839db9b1e3e11de0164f8e73a6ee08a6e434b4f2421df2"},"schema_version":"1.0","source":{"id":"2001.08764","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2001.08764","created_at":"2026-07-05T01:47:42Z"},{"alias_kind":"arxiv_version","alias_value":"2001.08764v2","created_at":"2026-07-05T01:47:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2001.08764","created_at":"2026-07-05T01:47:42Z"},{"alias_kind":"pith_short_12","alias_value":"OAD2UOTP2M4N","created_at":"2026-07-05T01:47:42Z"},{"alias_kind":"pith_short_16","alias_value":"OAD2UOTP2M4N7MM7","created_at":"2026-07-05T01:47:42Z"},{"alias_kind":"pith_short_8","alias_value":"OAD2UOTP","created_at":"2026-07-05T01:47:42Z"}],"graph_snapshots":[{"event_id":"sha256:c6265490d23db770ec0047de77bf17a90540a02e4bd24a39daec218f2ddece95","target":"graph","created_at":"2026-07-05T01:47:42Z","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/2001.08764/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large-scale, transformer-based language models such as GPT-2 are pretrained on diverse corpora scraped from the internet. Consequently, they are prone to generating non-normative text (i.e. in violation of social norms). We introduce a technique for fine-tuning GPT-2, using a policy gradient reinforcement learning technique and a normative text classifier to produce reward and punishment values. We evaluate our technique on five data sets using automated and human participant experiments. The normative text classifier is 81-90% accurate when compared to gold-standard human judgments of normati","authors_text":"Mark Riedl, Siyan Li, Spencer Frazier, Xiangyu Peng","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-01-23T19:06:18Z","title":"Reducing Non-Normative Text Generation from Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2001.08764","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:ddd1faba095b928cc211a15428a2e96f6dd87e4868309983f8cb37f2e323f9d0","target":"record","created_at":"2026-07-05T01:47:42Z","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":"319dd14409b45055ae6ce2bf6b110cde5118fd2a7a9a2b32d8030828ad213ba3","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-01-23T19:06:18Z","title_canon_sha256":"2a728d844524ce67f9839db9b1e3e11de0164f8e73a6ee08a6e434b4f2421df2"},"schema_version":"1.0","source":{"id":"2001.08764","kind":"arxiv","version":2}},"canonical_sha256":"7007aa3a6fd338dfb19f640e13300116c37e9a09437d78a17692f82606d1c4d5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7007aa3a6fd338dfb19f640e13300116c37e9a09437d78a17692f82606d1c4d5","first_computed_at":"2026-07-05T01:47:42.824121Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T01:47:42.824121Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"/D6935ZM/Zi1cLWbgQ+ALNxr+640nfNGY2fwlluof4gklo3u2mKjnN5aD45PmEIi3YJhsQM8w+u2hIrk0+6SCg==","signature_status":"signed_v1","signed_at":"2026-07-05T01:47:42.824522Z","signed_message":"canonical_sha256_bytes"},"source_id":"2001.08764","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ddd1faba095b928cc211a15428a2e96f6dd87e4868309983f8cb37f2e323f9d0","sha256:c6265490d23db770ec0047de77bf17a90540a02e4bd24a39daec218f2ddece95"],"state_sha256":"840f34f76bdf75169d4f951246ec85ed5509e694f5ba3dafb181ee0ce590f695"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5HkbLfj/qIJSthUDLr3wNEMtaw+pwtiZAdNDbvg2Bjh0/rxiodcOM6mJbZ7Pv/Vqu9n9266X8qwvSdtZ/CUGAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T10:20:10.848882Z","bundle_sha256":"45064d83547302658cad63aa0d3a7e430dbce1e3280157ac9432ad8e7a2b113c"}}