{"version":1,"type":"story","url":"https://digestai.news/story/research-shows-evidence-quality-boosts-factverification-scores","json":"https://digestai.news/story/research-shows-evidence-quality-boosts-factverification-scores.json","markdown":"https://digestai.news/story/research-shows-evidence-quality-boosts-factverification-scores.md","slug":"research-shows-evidence-quality-boosts-factverification-scores","headline":"Research shows evidence quality boosts fact‑verification scores","summary":"What Changes When Fact-Verification Scores Improve? Evidence and Answer Accounting Across Trained Verifiers and LLMs\narXiv:2609.27064v1 Announce Type: new \nAbstract: A joint fact-verification score assesses answers and submitted evidence together. When the score improves, how much of the gain remains if the answers are held fixed? On FEVEROUS, strict score is the percentage of claims with a correct answer and a complete annotated evidence group in the submitted evidence. Across four trained DeBERTa checkpoints and 7,890 claims, replacing DCUF evidence with UnifEE evidence raises strict score by 9.61 percentage points, compared with 1.96 percentage points in answer accuracy.","keyPoints":["Replacing DCUF evidence with UnifEE raises strict score by 9.61 pp."],"whyItMatters":"The findings highlight that improving evidence quality can substantially raise fact‑verification performance, guiding dataset curators and model developers to prioritize evidence curation over answer generation. They also show that larger context windows benefit evidence‑based scoring, informing future model architecture and training strategies for reliable verification systems.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":[],"models":["DeBERTa","Qwen","Llama"],"people":[]},"firstPublishedAt":"2026-09-24T04:00:00Z","updatedAt":"2026-09-24T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"arXiv cs.CL","title":"What Changes When Fact-Verification Scores Improve? Evidence and Answer Accounting Across Trained Verifiers and LLMs","url":"https://arxiv.org/abs/2609.27064","publishedAt":"2026-09-24T04:00:00Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Research shows evidence quality boosts fact‑verification scores\", 24 September 2026, https://digestai.news/story/research-shows-evidence-quality-boosts-factverification-scores","publisher":"Digest AI","title":"Research shows evidence quality boosts fact‑verification scores","datePublished":"2026-09-24T04:00:00Z","url":"https://digestai.news/story/research-shows-evidence-quality-boosts-factverification-scores"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}