← 研踪 PaperTrailarXiv:2611.04921 · v1
PROSE: Preservation-Constrained Selective Editing for Guilty Bias in Legal Judgment Prediction
首次提交 2026-11-03 21:58
版本更新 2026-11-03 21:58
雷达观测 2026-11-04 23:39
摘要
Legal judgment prediction models may retain a guilty verdict after a legally decisive condition has been altered such that the specified charge is no longer established. We study this guilty bias via selective repair over matched charge-established and charge-not-established cases, where the unsupported verdict should change while its supported counterpart remains stable. We construct CAIL-MC, a legally reviewed dataset of 11,478 matched pairs covering 261 Chinese criminal charges. Held-out activation interventions show that contrasts learned from correctly resolved pairs transfer to false-guilty cases, but target-reaching interventions can also damage matched supported-guilty decisions. Motivated by this observation, we introduce PROSE, which reconstructs case-specific targets from a training-derived residual geometry, aligns them with a local margin-descent signal, attenuates supported-guilty-sensitive components, and realizes bounded targets through a preservation-regularized low-rank parameter update. Across Qwen3-8B and GLM4-9B, PROSE achieves stronger pairwise safe repair and evidence sensitivity, while preserving standard single-pass inference.
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-03 21:58 · 最新观测版本
PROSE: Preservation-Constrained Selective Editing for Guilty Bias in Legal Judgment Prediction
Legal judgment prediction models may retain a guilty verdict after a legally decisive condition has been altered such that the specified charge is no longer established. We study this guilty bias via selective repair over matched charge-established and charge-not-established cases, where the unsupported verdict should change while its supported counterpart remains stable. We construct CAIL-MC, a legally reviewed dataset of 11,478 matched pairs covering 261 Chinese criminal charges. Held-out activation interventions show that contrasts learned from correctly resolved pairs transfer to false-guilty cases, but target-reaching interventions can also damage matched supported-guilty decisions. Motivated by this observation, we introduce PROSE, which reconstructs case-specific targets from a training-derived residual geometry, aligns them with a local margin-descent signal, attenuates supported-guilty-sensitive components, and realizes bounded targets through a preservation-regularized low-rank parameter update. Across Qwen3-8B and GLM4-9B, PROSE achieves stronger pairwise safe repair and evidence sensitivity, while preserving standard single-pass inference.
查看此版本解读 →查看此版本原文 ↗