Measuring & Mitigating Over-Alignment for LLMs in Multilingual Criminal Law Courts

• URL: http://arxiv.org/abs/2606.23375v2
• Date: Tue, 30 Jun 2026 12:22:05 GMT
• Title: Measuring & Mitigating Over-Alignment for LLMs in Multilingual Criminal Law Courts
• Abstract summary: The Swiss Federal Supreme Court uses small on-premises models for tentative translations and short-passage summarization across the four official languages.<n>We introduce TF-RefusalBench, a benchmark for criminal-law translation and summarization derived from public Swiss Supreme Court rulings.<n>We then use TF-RefusalBench to show that over-alignment is a multifaceted phenomenon, influenced by the model and the prompt and text languages being processed.
• Score: 0.25604314983611276
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: While the wider applicability of LLMs in the legal field is currently debated due to their reliability and the gravity of any errors, narrow uses with well-understood and mitigated risks have emerged. Notably the Swiss Federal Supreme Court uses small on-premises models for tentative translations and short-passage summarization across the four official languages. However, such usage is challenging in the context of Criminal Law. Since rulings and cases employees work on routinely can contain detailed descriptions of violent and sexual offenses, their legitimate work is compromised by refusals and disclaimers due to the activation of model guardrails (over-alignment). To measure this phenomenon, we introduce TF-RefusalBench, a multilingual benchmark for criminal-law translation and summarization derived from public Swiss Supreme Court rulings. TF-RefusalBench contains 5,200 total prompts across French, German, Italian, and English, corresponding to common task prompts and passages likely to trigger refusal. We then use TF-RefusalBench to show that over-alignment is a multifaceted phenomenon, influenced by the model and the prompt and text languages being processed, and that its impact cannot be evaluated solely from an over-refusal perspective, given the disclaimer's impact on task faithfulness. Finally, we evaluate approaches to enable on-premises LLMs for Criminal Law Tasks, demonstrating that while prompting can be effective, abliteration (refusal directions ablation) eliminates refusal with minimal impact on task performance.

Related papers

• Inference-Time Steering for Cross-Lingual Factual Consistency in LLMs [51.56484100374058]
  Large Language Models (LLMs) demonstrate remarkable multilingual fluency, but their internal knowledge representations remain disproportionately biased toward high-resource languages.<n>We investigate whether these biases can be mitigated at inference time, forcing an English-prompted model to answer as if it were queried in target languages.<n>Experiments on Gemma 3 12B Instruct reveal persona prompting to be the strongest overall intervention, balancing efficacy, safety, and out-of-domain generalization.
  arXiv  Detail & Related papers  (2026-07-21T16:15:05Z)
• WisdomInterrogatory (LuWen): An Open-Source Legal Large Language Model Technical Report [55.27414605169639]
  We present WisdomInterrogatory (LuWen), an open-source Chinese legal language model built upon the Baichuan foundation model.<n>We evaluate LuWen on five representative legal tasks spanning both prediction and generation settings.
  arXiv  Detail & Related papers  (2026-04-08T06:59:07Z)
• Analyzing Bias in False Refusal Behavior of Large Language Models for Hate Speech Detoxification [7.696781721646013]
  We investigate false refusal behavior in hate speech detoxification.<n>We show that large language models (LLMs) disproportionately refuse inputs with higher semantic toxicity.<n>We propose a simple cross-translation strategy, translating English hate speech into Chinese for detoxification and back.
  arXiv  Detail & Related papers  (2026-01-13T15:45:31Z)
• Large Language Models' Complicit Responses to Illicit Instructions across Socio-Legal Contexts [54.15982476754607]
  Large language models (LLMs) are now deployed at unprecedented scale, assisting millions of users in daily tasks.<n>This study defines complicit facilitation as the provision of guidance or support that enables illicit user instructions.<n>Using real-world legal cases and established legal frameworks, we construct an evaluation benchmark spanning 269 illicit scenarios and 50 illicit intents.
  arXiv  Detail & Related papers  (2025-11-25T16:01:31Z)
• Better Call CLAUSE: A Discrepancy Benchmark for Auditing LLMs Legal Reasoning Capabilities [15.35489310097019]
  CLAUSE is a first-of-its-kind benchmark designed to evaluate the fragility of an LLM's legal reasoning.<n>Our work outlines a path to identify and correct such reasoning failures in legal AI.
  arXiv  Detail & Related papers  (2025-11-01T00:51:21Z)
• Not ready for the bench: LLM legal interpretation is unstable and out of step with human judgments [2.8622281002418357]
  Recent scholarship has proposed that legal practitioners add large language models (LLMs) to their interpretive toolkit.<n>This work offers an empirical argument against LLM interpretation as recently practiced by legal scholars and federal judges.<n>Our investigation in English shows that models do not provide stable interpretive judgments.
  arXiv  Detail & Related papers  (2025-10-29T10:21:25Z)
• Universal Legal Article Prediction via Tight Collaboration between Supervised Classification Model and LLM [42.11889345473452]
  Legal Article Prediction (LAP) is a critical task in legal text classification.<n>We propose Uni-LAP, a universal framework for legal article prediction.
  arXiv  Detail & Related papers  (2025-09-26T09:42:20Z)
• Refusal Direction is Universal Across Safety-Aligned Languages [66.64709923081745]
  In this paper, we investigate the refusal behavior in large language models (LLMs) across 14 languages using PolyRefuse.<n>We uncover the surprising cross-lingual universality of the refusal direction: a vector extracted from English can bypass refusals in other languages with near-perfect effectiveness.<n>We attribute this transferability to the parallelism of refusal vectors across languages in the embedding space and identify the underlying mechanism behind cross-lingual jailbreaks.
  arXiv  Detail & Related papers  (2025-05-22T21:54:46Z)
• DELTA: Pre-train a Discriminative Encoder for Legal Case Retrieval via Structural Word Alignment [55.91429725404988]
  We introduce DELTA, a discriminative model designed for legal case retrieval. We leverage shallow decoders to create information bottlenecks, aiming to enhance the representation ability. Our approach can outperform existing state-of-the-art methods in legal case retrieval.
  arXiv  Detail & Related papers  (2024-03-27T10:40:14Z)
• Multilingual Jailbreak Challenges in Large Language Models [96.74878032417054]
  In this study, we reveal the presence of multilingual jailbreak challenges within large language models (LLMs) We consider two potential risky scenarios: unintentional and intentional. We propose a novel textscSelf-Defense framework that automatically generates multilingual training data for safety fine-tuning.
  arXiv  Detail & Related papers  (2023-10-10T09:44:06Z)
• A Legal Approach to Hate Speech: Operationalizing the EU's Legal Framework against the Expression of Hatred as an NLP Task [2.248133901806859]
  We propose a 'legal approach' to hate speech detection by operationalization of the decision as to whether a post is subject to criminal law. We show that, by breaking the legal assessment down into a series of simpler sub-decisions, even laypersons can annotate consistently.
  arXiv  Detail & Related papers  (2020-04-07T14:13:38Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.