Skip to main content
Search

# Legal Judgment Prediction

2 related articles

Subscribe to updates
  1. arXiv

    Legal judgment models keep guilty verdicts after the decisive fact changes: PROSE cuts damage to supported rulings from 17.5% to 8.5%

    Legal judgment prediction models often keep a guilty verdict after a legally decisive condition has been altered, a tendency the authors call guilty bias. PROSE frames the correction as a paired constraint: the unsupported member of a case pair should flip, while its counterpart, differing in one decisive condition, should stay. On Qwen3-8B and GLM-4-9B it performs this selective repair through a one-time offline parameter edit, and at matched repair rates it reduces damage to supported verdicts from 17.5% to 8.5% on Qwen and from 24.3% to 7.1% on GLM.

  2. Three agents argue a Supreme Court case: precedent retrieval lifts criminal petitioner Legal-sim to 0.8921

    Judgment prediction has mostly reasoned over case facts, yet real proceedings are shaped by exchanges among judge, petitioner and respondent. ArguGenCourt assigns three independent GPT-4o agents these roles and lets them converse in rounds, seeding prompts with courtroom utterances retrieved from the factually closest precedent document. On criminal petitioner arguments, local precedent retrieval in the multi-agent setting reaches a Legal-sim of 0.8921 against 0.8288 for fact-based conversation under the same retrieval; feeding the generated arguments into judgment prediction gives 0.55 accuracy versus 0.48 for zero-shot.

You’re all caught up