hidden-representations

本文介绍了一种通过分析探针轨迹（即概念概率在生成token上的演变）来监控大型推理模型推理过程的方法。该方法利用隐藏表示中的时间特征和信号处理特征，更好地预测未来模型行为，通过最大池化达到了高达95%的AUROC。

This paper demonstrates that a small post-transformer adapter (786K parameters) can correct suppressed log-probabilities in alignment-tuned language models, particularly on politically sensitive topics. The adapter shows 31-39% generalization to held-out facts across Qwen3 models while maintaining coherent generation when applied at the final prediction position.