发表机构
Microsoft India（微软印度）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究揭示扩散语言模型存在表示-置信度差距，其高置信度集中是误导性症状，现有补救措施难修复排序缺陷，提出轻量级工具利用隐藏状态信号改进排序，指出噪声下置信度可靠性是更紧迫限制。
AI 中文摘要
扩散语言模型利用广泛的上下文生成文本，表明它们可能比标准模型更能处理输入噪声。测试显示这仅部分正确：扩散模型内部能高度准确地检测文本错误，但其外部报告的置信度却忽略了这一信号。当因噪声导致准确率下降时，置信度仍保持在接近最大值，且正确排序答案的能力会退化为随机水平，我们将这种不匹配称为表示-置信度差距。高置信度分数的明显集中是一种误导性的表面症状，标准数学调整可消除这种集中，但无法修复底层的排序顺序损失。这种排序缺陷使标准模型在噪声条件下更具优势，且难以通过常见补救措施解决：匹配训练可恢复准确率但无法恢复排序，分数重新校准和输入级错误信号也无法重新排序最终答案。不过，评估答案所需的信息仍存在于隐藏状态中，一种轻量级提取工具利用该信号改进排序，该方法效率极高，因为它完全冻结基础模型，且无需额外的文本生成步骤。我们推出该工具以证明该信号存在，同时明确指出其局限性。最终，在噪声条件下，置信度可靠性比整体准确率是更紧迫的限制因素。
英文摘要
Diffusion language models use broad context to create text, suggesting they might handle input noise better than standard models. Testing reveals this is only partially true. Internally, diffusion models detect text errors highly accurately. Externally, their reported certainty ignores this signal. As accuracy drops due to noise, confidence stays near its maximum and the ability to correctly rank answers degrades toward random chance. We call this mismatch the representation confidence gap. The visible concentration of high certainty scores is a misleading surface symptom. Standard math adjustments remove this concentration but fail to fix the underlying loss of ranking order. This ranking deficit favors standard models under noisy conditions and resists common remedies. Matching training recovers accuracy but not ranking, while score recalibration and input level error signals cannot reorder the final answers. However, the information needed to properly evaluate an answer survives in the hidden states. A lightweight extraction tool uses this signal to improve ranking. This approach is highly efficient because it leaves the base model completely frozen and requires zero additional text generation steps. We present this tool to prove the signal exists, while clearly noting its limits. Ultimately, certainty reliability is a more pressing limit than overall accuracy under noisy conditions.
