论文提出 J-ICL：让大模型学会判断上下文该不该相信
When Context Misleads: In-context Learning with Jurisdiction in Large Language Models
这篇论文发现 ICL 微调会让模型更容易被假上下文带偏，实测最多降 14.95 个点，J-ICL 方法能两边兼顾，做 RAG 或智能体的朋友可以看看。
arXiv 论文指出，现有 In-Context Learning 微调方法只顾从示例中提取模式，忽略了判断上下文是否可信的"上下文权威"能力。作者构建了 FakeContextBench 基准，覆盖七个领域的伪科学声明，发现常见 ICL 微调会让模型更易被误导，现实准确率最多比基座模型下降 14.95 个百分点。论文提出的 J-ICL 框架把上下文验证纳入训练目标，在四个模型骨干上平均将 ICLEval 提升 5.84 个百分点，现实准确率提升 9.20 个百分点。相比 MetaICL 和 Symbol Tuning，J-ICL 的 Reality Rate 平均高出 18.09 个百分点。基准已开源在 GitHub 上。
When Context Misleads: In-context Learning with Jurisdiction in Large Language Models
In-Context Learning (ICL) has become a cornerstone of modern LLM deployment. However, existing ICL post-training methods have a critical blind spot: they excel at extracting patterns from demonstrations while often neglecting context authority, the ability to determine whether contextual information should govern the final answer. To benchmark this capability, we introduce FakeContextBench, which contains pseudoscientific claims across seven domains. Our evaluation of commercial and open-source models shows that large-scale pre-training alone is insufficient for reliable context-authority discrimination. Moreover, prevalent ICL fine-tuning methods can increase susceptibility to misleading context, reducing reality accuracy by up to 14.95 percentage points relative to the base model. To address this trade-off, we propose Jurisdiction In-Context Learning (J-ICL), a post-training framework that incorporates context validation into the training objective. Across four model backbones, J-ICL improves ICLEval by an average of 5.84 percentage points and reality accuracy by 9.20 points over the corresponding base models. It also raises the Reality Rate by an average of 18.09 points relative to MetaICL and Symbol Tuning. These results demonstrate that ICL capability and resistance to deceptive context can be improved together. The benchmark is available at https://github.com/peilin717/FakeContext-Bench.
