{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:F4G3NPNQHBDGIS4EIBY265E52O","short_pith_number":"pith:F4G3NPNQ","canonical_record":{"source":{"id":"2605.13709","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T15:56:37Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"4eabab187baddfd8d928f3847a3196b5331f46c48e2c251014d5649f4e0a31fe","abstract_canon_sha256":"eb1a0dcf05cd4e12868a86f720c5105972940b3006ffe4a25e9e7463c09b77c9"},"schema_version":"1.0"},"canonical_sha256":"2f0db6bdb03846644b844071af749dd387d5795f4709ad579b6ffde97a6ca50c","source":{"kind":"arxiv","id":"2605.13709","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13709","created_at":"2026-05-18T02:44:16Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13709v1","created_at":"2026-05-18T02:44:16Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13709","created_at":"2026-05-18T02:44:16Z"},{"alias_kind":"pith_short_12","alias_value":"F4G3NPNQHBDG","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"F4G3NPNQHBDGIS4E","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"F4G3NPNQ","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:F4G3NPNQHBDGIS4EIBY265E52O","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13709","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T15:56:37Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"4eabab187baddfd8d928f3847a3196b5331f46c48e2c251014d5649f4e0a31fe","abstract_canon_sha256":"eb1a0dcf05cd4e12868a86f720c5105972940b3006ffe4a25e9e7463c09b77c9"},"schema_version":"1.0"},"canonical_sha256":"2f0db6bdb03846644b844071af749dd387d5795f4709ad579b6ffde97a6ca50c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:16.774928Z","signature_b64":"NiyMG/MDo2PiVpiSQKLsnJCOCbJCBoPqVmNgOSNSVgzUha8MFB6C5k16+APdM6Meux3M4eLuZ25ANF7GuLOTAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2f0db6bdb03846644b844071af749dd387d5795f4709ad579b6ffde97a6ca50c","last_reissued_at":"2026-05-18T02:44:16.774397Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:16.774397Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13709","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:16Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"l46dw8Xlo9Jkl2X+S99LAL4VjXFyHN4nUhNWF0fiDS5hIU98AMIvqdU1sLLbritUL5nbMLE1sZUWioC6GJR1Cw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T05:05:55.781789Z"},"content_sha256":"72bb6080993e6c8ed4c3b517fa25e606164d63816808622e9b2d749e95ab9fd7","schema_version":"1.0","event_id":"sha256:72bb6080993e6c8ed4c3b517fa25e606164d63816808622e9b2d749e95ab9fd7"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:F4G3NPNQHBDGIS4EIBY265E52O","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Children's English Reading Story Generation via Supervised Fine-Tuning of Compact LLMs with Controllable Difficulty and Safety","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Fine-tuned 8B LLMs generate children's reading stories that better match target difficulty levels than zero-shot outputs from GPT-4o or Llama 3.3 70B.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bonnie J. Dorr (1), Fanghua Cao (1), Gainesville, Min Yao (1), Qian Shen (1), Shlok Gilda (1), USA), Walter L. Leite (1) ((1) University of Florida","submitted_at":"2026-05-13T15:56:37Z","abstract_excerpt":"Large Language Models (LLMs) are widely applied in educational practices, such as for generating children's stories. However, the generated stories are often too difficult for children to read, and the operational cost of LLMs hinders their widespread adoption in educational settings. We used an existing expert-designed children's reading curriculum and its corresponding generated stories from GPT-4o and Llama 3.3 70B to design different experiments for fine-tuning three 8B-parameter LLMs, which then generated new English reading stories that were subjected to quantitative and qualitative eval"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our evaluation results show that with appropriate fine-tuning designs, children's English reading stories generated by 8B LLMs perform better on difficulty-related metrics than those from zero-shot GPT-4o and Llama 3.3 70B, with almost no discernible safety issues.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the chosen quantitative difficulty metrics and qualitative safety checks accurately capture real-world readability and safety for children, and that fine-tuning on GPT-generated stories does not introduce hidden biases or reduce engagement.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Fine-tuned 8B LLMs produce children's English reading stories with better difficulty control and safety than zero-shot GPT-4o and Llama 3.3 70B.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Fine-tuned 8B LLMs generate children's reading stories that better match target difficulty levels than zero-shot outputs from GPT-4o or Llama 3.3 70B.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9a65fb0f13c293f33458757f0b21d603654d24c34c361a6a4c366abdd2f30b7d"},"source":{"id":"2605.13709","kind":"arxiv","version":1},"verdict":{"id":"6b60a85e-eb57-4ddd-8616-57d773362a2b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:28:28.671467Z","strongest_claim":"Our evaluation results show that with appropriate fine-tuning designs, children's English reading stories generated by 8B LLMs perform better on difficulty-related metrics than those from zero-shot GPT-4o and Llama 3.3 70B, with almost no discernible safety issues.","one_line_summary":"Fine-tuned 8B LLMs produce children's English reading stories with better difficulty control and safety than zero-shot GPT-4o and Llama 3.3 70B.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the chosen quantitative difficulty metrics and qualitative safety checks accurately capture real-world readability and safety for children, and that fine-tuning on GPT-generated stories does not introduce hidden biases or reduce engagement.","pith_extraction_headline":"Fine-tuned 8B LLMs generate children's reading stories that better match target difficulty levels than zero-shot outputs from GPT-4o or Llama 3.3 70B."},"references":{"count":49,"sample":[{"doi":"","year":2025,"title":"Journal of Children and Media , volume=","work_id":"31f98e3d-8e06-487a-a274-735b5b53ca8f","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2006,"title":"Reading and Writing , volume=","work_id":"78b2a5a6-b862-4243-9532-9316ca4fdf5f","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2015,"title":"Journal of Research in Reading , volume=","work_id":"e44c9ee0-ae11-4b61-b92e-35bc647ee892","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2014,"title":"Procedia-Social and Behavioral Sciences , volume=","work_id":"75e40be2-520e-4801-8750-c2859cf57baf","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Generative AI and ChatGPT in school children’s education: Evidence from a school lesson , author=. Sustainability , volume=. 2023 , publisher=","work_id":"481256eb-11d9-4d3d-801d-67841c6299a3","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":49,"snapshot_sha256":"e32bc71631cc6df952a8fa542897ec194c24301067c028370ae982e8e2db4a52","internal_anchors":2},"formal_canon":{"evidence_count":2,"snapshot_sha256":"48cffc9f52fd42069cd24d48fc3ed923706b15971910e550fe11788a3d2c9b6b"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"6b60a85e-eb57-4ddd-8616-57d773362a2b"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:16Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"UTC2bwDrYjV68edXj1nGYGF3X4I9u1aqtcZyzGkHzjrvEJqlqKpqSBp8P5ObDNCUvRDMShEAdPz/mDauqyzXCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T05:05:55.783009Z"},"content_sha256":"04cd607f77cd3fe607818da6ab6d445935dcff5bcb2399d8443f5efca5362c7b","schema_version":"1.0","event_id":"sha256:04cd607f77cd3fe607818da6ab6d445935dcff5bcb2399d8443f5efca5362c7b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/F4G3NPNQHBDGIS4EIBY265E52O/bundle.json","state_url":"https://pith.science/pith/F4G3NPNQHBDGIS4EIBY265E52O/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/F4G3NPNQHBDGIS4EIBY265E52O/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":"2026-05-25T05:05:55Z","links":{"resolver":"https://pith.science/pith/F4G3NPNQHBDGIS4EIBY265E52O","bundle":"https://pith.science/pith/F4G3NPNQHBDGIS4EIBY265E52O/bundle.json","state":"https://pith.science/pith/F4G3NPNQHBDGIS4EIBY265E52O/state.json","well_known_bundle":"https://pith.science/.well-known/pith/F4G3NPNQHBDGIS4EIBY265E52O/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:F4G3NPNQHBDGIS4EIBY265E52O","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":"eb1a0dcf05cd4e12868a86f720c5105972940b3006ffe4a25e9e7463c09b77c9","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T15:56:37Z","title_canon_sha256":"4eabab187baddfd8d928f3847a3196b5331f46c48e2c251014d5649f4e0a31fe"},"schema_version":"1.0","source":{"id":"2605.13709","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13709","created_at":"2026-05-18T02:44:16Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13709v1","created_at":"2026-05-18T02:44:16Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13709","created_at":"2026-05-18T02:44:16Z"},{"alias_kind":"pith_short_12","alias_value":"F4G3NPNQHBDG","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"F4G3NPNQHBDGIS4E","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"F4G3NPNQ","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:04cd607f77cd3fe607818da6ab6d445935dcff5bcb2399d8443f5efca5362c7b","target":"graph","created_at":"2026-05-18T02:44:16Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Our evaluation results show that with appropriate fine-tuning designs, children's English reading stories generated by 8B LLMs perform better on difficulty-related metrics than those from zero-shot GPT-4o and Llama 3.3 70B, with almost no discernible safety issues."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the chosen quantitative difficulty metrics and qualitative safety checks accurately capture real-world readability and safety for children, and that fine-tuning on GPT-generated stories does not introduce hidden biases or reduce engagement."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Fine-tuned 8B LLMs produce children's English reading stories with better difficulty control and safety than zero-shot GPT-4o and Llama 3.3 70B."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Fine-tuned 8B LLMs generate children's reading stories that better match target difficulty levels than zero-shot outputs from GPT-4o or Llama 3.3 70B."}],"snapshot_sha256":"9a65fb0f13c293f33458757f0b21d603654d24c34c361a6a4c366abdd2f30b7d"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"48cffc9f52fd42069cd24d48fc3ed923706b15971910e550fe11788a3d2c9b6b"},"paper":{"abstract_excerpt":"Large Language Models (LLMs) are widely applied in educational practices, such as for generating children's stories. However, the generated stories are often too difficult for children to read, and the operational cost of LLMs hinders their widespread adoption in educational settings. We used an existing expert-designed children's reading curriculum and its corresponding generated stories from GPT-4o and Llama 3.3 70B to design different experiments for fine-tuning three 8B-parameter LLMs, which then generated new English reading stories that were subjected to quantitative and qualitative eval","authors_text":"Bonnie J. Dorr (1), Fanghua Cao (1), Gainesville, Min Yao (1), Qian Shen (1), Shlok Gilda (1), USA), Walter L. Leite (1) ((1) University of Florida","cross_cats":["cs.AI","cs.LG"],"headline":"Fine-tuned 8B LLMs generate children's reading stories that better match target difficulty levels than zero-shot outputs from GPT-4o or Llama 3.3 70B.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T15:56:37Z","title":"Children's English Reading Story Generation via Supervised Fine-Tuning of Compact LLMs with Controllable Difficulty and Safety"},"references":{"count":49,"internal_anchors":2,"resolved_work":49,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Journal of Children and Media , volume=","work_id":"31f98e3d-8e06-487a-a274-735b5b53ca8f","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Reading and Writing , volume=","work_id":"78b2a5a6-b862-4243-9532-9316ca4fdf5f","year":2006},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Journal of Research in Reading , volume=","work_id":"e44c9ee0-ae11-4b61-b92e-35bc647ee892","year":2015},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Procedia-Social and Behavioral Sciences , volume=","work_id":"75e40be2-520e-4801-8750-c2859cf57baf","year":2014},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Generative AI and ChatGPT in school children’s education: Evidence from a school lesson , author=. Sustainability , volume=. 2023 , publisher=","work_id":"481256eb-11d9-4d3d-801d-67841c6299a3","year":2023}],"snapshot_sha256":"e32bc71631cc6df952a8fa542897ec194c24301067c028370ae982e8e2db4a52"},"source":{"id":"2605.13709","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T19:28:28.671467Z","id":"6b60a85e-eb57-4ddd-8616-57d773362a2b","model_set":{"reader":"grok-4.3"},"one_line_summary":"Fine-tuned 8B LLMs produce children's English reading stories with better difficulty control and safety than zero-shot GPT-4o and Llama 3.3 70B.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Fine-tuned 8B LLMs generate children's reading stories that better match target difficulty levels than zero-shot outputs from GPT-4o or Llama 3.3 70B.","strongest_claim":"Our evaluation results show that with appropriate fine-tuning designs, children's English reading stories generated by 8B LLMs perform better on difficulty-related metrics than those from zero-shot GPT-4o and Llama 3.3 70B, with almost no discernible safety issues.","weakest_assumption":"That the chosen quantitative difficulty metrics and qualitative safety checks accurately capture real-world readability and safety for children, and that fine-tuning on GPT-generated stories does not introduce hidden biases or reduce engagement."}},"verdict_id":"6b60a85e-eb57-4ddd-8616-57d773362a2b"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:72bb6080993e6c8ed4c3b517fa25e606164d63816808622e9b2d749e95ab9fd7","target":"record","created_at":"2026-05-18T02:44:16Z","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":"eb1a0dcf05cd4e12868a86f720c5105972940b3006ffe4a25e9e7463c09b77c9","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T15:56:37Z","title_canon_sha256":"4eabab187baddfd8d928f3847a3196b5331f46c48e2c251014d5649f4e0a31fe"},"schema_version":"1.0","source":{"id":"2605.13709","kind":"arxiv","version":1}},"canonical_sha256":"2f0db6bdb03846644b844071af749dd387d5795f4709ad579b6ffde97a6ca50c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2f0db6bdb03846644b844071af749dd387d5795f4709ad579b6ffde97a6ca50c","first_computed_at":"2026-05-18T02:44:16.774397Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:16.774397Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"NiyMG/MDo2PiVpiSQKLsnJCOCbJCBoPqVmNgOSNSVgzUha8MFB6C5k16+APdM6Meux3M4eLuZ25ANF7GuLOTAw==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:16.774928Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13709","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:72bb6080993e6c8ed4c3b517fa25e606164d63816808622e9b2d749e95ab9fd7","sha256:04cd607f77cd3fe607818da6ab6d445935dcff5bcb2399d8443f5efca5362c7b"],"state_sha256":"88b60a32487145b91942066b806f74f76392653388ea1166946bba0fe736a8a1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DUYKQr6l7636MOPxzP7N32nVW4F2PTtl3yCk6n8gg0OyJCYHAqVk06uRDYCy0LxXsmFWbbkRSZqNH/Pxn9GjAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-25T05:05:55.788786Z","bundle_sha256":"c0358c55d214aceaea4b87aab2988a1c0b4d7fe78b2fb12ca74ab80cca637c7d"}}