{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:5QKHNOFAGOPSDUNNGKVP3H5MZP","short_pith_number":"pith:5QKHNOFA","canonical_record":{"source":{"id":"2309.12546","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-09-22T00:13:07Z","cross_cats_sorted":[],"title_canon_sha256":"4f22b59b29531194e216406845825956fc39f4a38ae4926da99da9cea876f417","abstract_canon_sha256":"74dec409e99753b5af875a10815a60c1836f66a132df77be2bb47261eddc8176"},"schema_version":"1.0"},"canonical_sha256":"ec1476b8a0339f21d1ad32aafd9faccbd71cc87718d8f600ab38368a8e554410","source":{"kind":"arxiv","id":"2309.12546","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2309.12546","created_at":"2026-07-05T07:49:14Z"},{"alias_kind":"arxiv_version","alias_value":"2309.12546v2","created_at":"2026-07-05T07:49:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.12546","created_at":"2026-07-05T07:49:14Z"},{"alias_kind":"pith_short_12","alias_value":"5QKHNOFAGOPS","created_at":"2026-07-05T07:49:14Z"},{"alias_kind":"pith_short_16","alias_value":"5QKHNOFAGOPSDUNN","created_at":"2026-07-05T07:49:14Z"},{"alias_kind":"pith_short_8","alias_value":"5QKHNOFA","created_at":"2026-07-05T07:49:14Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:5QKHNOFAGOPSDUNNGKVP3H5MZP","target":"record","payload":{"canonical_record":{"source":{"id":"2309.12546","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-09-22T00:13:07Z","cross_cats_sorted":[],"title_canon_sha256":"4f22b59b29531194e216406845825956fc39f4a38ae4926da99da9cea876f417","abstract_canon_sha256":"74dec409e99753b5af875a10815a60c1836f66a132df77be2bb47261eddc8176"},"schema_version":"1.0"},"canonical_sha256":"ec1476b8a0339f21d1ad32aafd9faccbd71cc87718d8f600ab38368a8e554410","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:49:14.800981Z","signature_b64":"WjezQuK5zMbNyy+4ohneTmXmotj3qKpH/3sn3SusixwX7oRyrYj4foFXUEPzU+P6p6EPeke/mO+CNnnRsrQ1Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ec1476b8a0339f21d1ad32aafd9faccbd71cc87718d8f600ab38368a8e554410","last_reissued_at":"2026-07-05T07:49:14.800468Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:49:14.800468Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2309.12546","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":"2026-07-05T07:49:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"aKgYSxFAKQUhBNunJ4aK5XxlutBoxH83YjjxiaS6rR6LF6lQIqj+mUhy85UacIMkhZ5XF2iWB8I6ZuBOe6y9BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T01:19:30.592112Z"},"content_sha256":"9b77308c8d92ffddc1dfcf8323f19ebb895082ae0f01a849b8cf08d9a7d84d9c","schema_version":"1.0","event_id":"sha256:9b77308c8d92ffddc1dfcf8323f19ebb895082ae0f01a849b8cf08d9a7d84d9c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:5QKHNOFAGOPSDUNNGKVP3H5MZP","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Automatic Answerability Evaluation for Question Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Kotaro Funakoshi, Manabu Okumura, Zifan Wang","submitted_at":"2023-09-22T00:13:07Z","abstract_excerpt":"Conventional automatic evaluation metrics, such as BLEU and ROUGE, developed for natural language generation (NLG) tasks, are based on measuring the n-gram overlap between the generated and reference text. These simple metrics may be insufficient for more complex tasks, such as question generation (QG), which requires generating questions that are answerable by the reference answers. Developing a more sophisticated automatic evaluation metric, thus, remains an urgent problem in QG research. This work proposes PMAN (Prompting-based Metric on ANswerability), a novel automatic evaluation metric t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.12546","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/2309.12546/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":"2026-07-05T07:49:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wIlq9+1gTm1ZZW3DzmP/5za7O0WGvjyRoPAZLwQu17pSJU7TP78phQ3w2ux5P8hwF9AoKB7vSVN6TjX7zFy4Aw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T01:19:30.592612Z"},"content_sha256":"829261721c72a047a1c908bd7d1628ba94fbff31b3b2f84d60af86a5f51d44b2","schema_version":"1.0","event_id":"sha256:829261721c72a047a1c908bd7d1628ba94fbff31b3b2f84d60af86a5f51d44b2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5QKHNOFAGOPSDUNNGKVP3H5MZP/bundle.json","state_url":"https://pith.science/pith/5QKHNOFAGOPSDUNNGKVP3H5MZP/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5QKHNOFAGOPSDUNNGKVP3H5MZP/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":"2026-08-03T01:19:30Z","links":{"resolver":"https://pith.science/pith/5QKHNOFAGOPSDUNNGKVP3H5MZP","bundle":"https://pith.science/pith/5QKHNOFAGOPSDUNNGKVP3H5MZP/bundle.json","state":"https://pith.science/pith/5QKHNOFAGOPSDUNNGKVP3H5MZP/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5QKHNOFAGOPSDUNNGKVP3H5MZP/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:5QKHNOFAGOPSDUNNGKVP3H5MZP","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":"74dec409e99753b5af875a10815a60c1836f66a132df77be2bb47261eddc8176","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-09-22T00:13:07Z","title_canon_sha256":"4f22b59b29531194e216406845825956fc39f4a38ae4926da99da9cea876f417"},"schema_version":"1.0","source":{"id":"2309.12546","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2309.12546","created_at":"2026-07-05T07:49:14Z"},{"alias_kind":"arxiv_version","alias_value":"2309.12546v2","created_at":"2026-07-05T07:49:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.12546","created_at":"2026-07-05T07:49:14Z"},{"alias_kind":"pith_short_12","alias_value":"5QKHNOFAGOPS","created_at":"2026-07-05T07:49:14Z"},{"alias_kind":"pith_short_16","alias_value":"5QKHNOFAGOPSDUNN","created_at":"2026-07-05T07:49:14Z"},{"alias_kind":"pith_short_8","alias_value":"5QKHNOFA","created_at":"2026-07-05T07:49:14Z"}],"graph_snapshots":[{"event_id":"sha256:829261721c72a047a1c908bd7d1628ba94fbff31b3b2f84d60af86a5f51d44b2","target":"graph","created_at":"2026-07-05T07:49:14Z","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/2309.12546/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Conventional automatic evaluation metrics, such as BLEU and ROUGE, developed for natural language generation (NLG) tasks, are based on measuring the n-gram overlap between the generated and reference text. These simple metrics may be insufficient for more complex tasks, such as question generation (QG), which requires generating questions that are answerable by the reference answers. Developing a more sophisticated automatic evaluation metric, thus, remains an urgent problem in QG research. This work proposes PMAN (Prompting-based Metric on ANswerability), a novel automatic evaluation metric t","authors_text":"Kotaro Funakoshi, Manabu Okumura, Zifan Wang","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-09-22T00:13:07Z","title":"Automatic Answerability Evaluation for Question Generation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.12546","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":"sha256:9b77308c8d92ffddc1dfcf8323f19ebb895082ae0f01a849b8cf08d9a7d84d9c","target":"record","created_at":"2026-07-05T07:49:14Z","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":"74dec409e99753b5af875a10815a60c1836f66a132df77be2bb47261eddc8176","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-09-22T00:13:07Z","title_canon_sha256":"4f22b59b29531194e216406845825956fc39f4a38ae4926da99da9cea876f417"},"schema_version":"1.0","source":{"id":"2309.12546","kind":"arxiv","version":2}},"canonical_sha256":"ec1476b8a0339f21d1ad32aafd9faccbd71cc87718d8f600ab38368a8e554410","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ec1476b8a0339f21d1ad32aafd9faccbd71cc87718d8f600ab38368a8e554410","first_computed_at":"2026-07-05T07:49:14.800468Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:49:14.800468Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"WjezQuK5zMbNyy+4ohneTmXmotj3qKpH/3sn3SusixwX7oRyrYj4foFXUEPzU+P6p6EPeke/mO+CNnnRsrQ1Bg==","signature_status":"signed_v1","signed_at":"2026-07-05T07:49:14.800981Z","signed_message":"canonical_sha256_bytes"},"source_id":"2309.12546","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9b77308c8d92ffddc1dfcf8323f19ebb895082ae0f01a849b8cf08d9a7d84d9c","sha256:829261721c72a047a1c908bd7d1628ba94fbff31b3b2f84d60af86a5f51d44b2"],"state_sha256":"17de4daa9efa622c449077e327468ea6987d7915d43c385053140f0fdb6cefde"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"GtwtrExGkHANxCRKeCF0nAmtVkva0b3+oVTxI9af4qFyn53SJ/BPsKPF0XH0rSCH7FD/XABbF0WhXnHJ4sP8Dg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-03T01:19:30.596150Z","bundle_sha256":"53ffa9a3ce151d6421f4e974bac3a29895899d59422dc5025ce709c9eb306c57"}}