{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:6HIT5EG5YA5IGQ6EWOALHDAC5T","short_pith_number":"pith:6HIT5EG5","schema_version":"1.0","canonical_sha256":"f1d13e90ddc03a8343c4b380b38c02ecdfd0cb34a5024423c68e380558da7c88","source":{"kind":"arxiv","id":"2312.15842","version":3},"attestation_state":"computed","paper":{"title":"Knowledge Distillation of LLM for Automatic Scoring of Science Education Assessments","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Ehsan Latif, Luyang Fang, Ping Ma, Xiaoming Zhai","submitted_at":"2023-12-26T01:24:25Z","abstract_excerpt":"This study proposes a method for knowledge distillation (KD) of fine-tuned Large Language Models (LLMs) into smaller, more efficient, and accurate neural networks. We specifically target the challenge of deploying these models on resource-constrained devices. Our methodology involves training the smaller student model (Neural Network) using the prediction probabilities (as soft labels) of the LLM, which serves as a teacher model. This is achieved through a specialized loss function tailored to learn from the LLM's output probabilities, ensuring that the student model closely mimics the teacher"},"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":"2312.15842","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-12-26T01:24:25Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9b8da0ce476a12fbfaaeb76c29b8dafd5696214fca541545d9221d0bb52ca6f1","abstract_canon_sha256":"4fa7de7a527479ee9a633720c5c7daedfe964bf5bcd651a28945578d333ad66c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:30:30.544752Z","signature_b64":"+5Jp7OH2ntgzRANxqBwjMuMzkUxjV3m3aL3szr7+wI6ZI0e/1YADyHpwH4tMxaPF1EqZq7TIn4UAObcF2Q30DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f1d13e90ddc03a8343c4b380b38c02ecdfd0cb34a5024423c68e380558da7c88","last_reissued_at":"2026-07-05T08:30:30.544217Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:30:30.544217Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Knowledge Distillation of LLM for Automatic Scoring of Science Education Assessments","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Ehsan Latif, Luyang Fang, Ping Ma, Xiaoming Zhai","submitted_at":"2023-12-26T01:24:25Z","abstract_excerpt":"This study proposes a method for knowledge distillation (KD) of fine-tuned Large Language Models (LLMs) into smaller, more efficient, and accurate neural networks. We specifically target the challenge of deploying these models on resource-constrained devices. Our methodology involves training the smaller student model (Neural Network) using the prediction probabilities (as soft labels) of the LLM, which serves as a teacher model. This is achieved through a specialized loss function tailored to learn from the LLM's output probabilities, ensuring that the student model closely mimics the teacher"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.15842","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/2312.15842/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":"2312.15842","created_at":"2026-07-05T08:30:30.544285+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.15842v3","created_at":"2026-07-05T08:30:30.544285+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.15842","created_at":"2026-07-05T08:30:30.544285+00:00"},{"alias_kind":"pith_short_12","alias_value":"6HIT5EG5YA5I","created_at":"2026-07-05T08:30:30.544285+00:00"},{"alias_kind":"pith_short_16","alias_value":"6HIT5EG5YA5IGQ6E","created_at":"2026-07-05T08:30:30.544285+00:00"},{"alias_kind":"pith_short_8","alias_value":"6HIT5EG5","created_at":"2026-07-05T08:30:30.544285+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24973","citing_title":"LLM Performance on a Real, Double-Marked GCSE Benchmark","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27967","citing_title":"Multi-Teacher Knowledge Distillation via Teacher-Informed Mixture Priors","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6HIT5EG5YA5IGQ6EWOALHDAC5T","json":"https://pith.science/pith/6HIT5EG5YA5IGQ6EWOALHDAC5T.json","graph_json":"https://pith.science/api/pith-number/6HIT5EG5YA5IGQ6EWOALHDAC5T/graph.json","events_json":"https://pith.science/api/pith-number/6HIT5EG5YA5IGQ6EWOALHDAC5T/events.json","paper":"https://pith.science/paper/6HIT5EG5"},"agent_actions":{"view_html":"https://pith.science/pith/6HIT5EG5YA5IGQ6EWOALHDAC5T","download_json":"https://pith.science/pith/6HIT5EG5YA5IGQ6EWOALHDAC5T.json","view_paper":"https://pith.science/paper/6HIT5EG5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.15842&json=true","fetch_graph":"https://pith.science/api/pith-number/6HIT5EG5YA5IGQ6EWOALHDAC5T/graph.json","fetch_events":"https://pith.science/api/pith-number/6HIT5EG5YA5IGQ6EWOALHDAC5T/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6HIT5EG5YA5IGQ6EWOALHDAC5T/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6HIT5EG5YA5IGQ6EWOALHDAC5T/action/storage_attestation","attest_author":"https://pith.science/pith/6HIT5EG5YA5IGQ6EWOALHDAC5T/action/author_attestation","sign_citation":"https://pith.science/pith/6HIT5EG5YA5IGQ6EWOALHDAC5T/action/citation_signature","submit_replication":"https://pith.science/pith/6HIT5EG5YA5IGQ6EWOALHDAC5T/action/replication_record"}},"created_at":"2026-07-05T08:30:30.544285+00:00","updated_at":"2026-07-05T08:30:30.544285+00:00"}