Chain-of-Thought Faithfulness of Reasoning Models Varies with Where and How Preference Cues Are Delivered
研究通过FACE-Eval评估了不同位置和形式的偏好提示对推理模型忠实度的影响，发现工具返回和隐式提示降低了模型的明确承诺度，增加了未表达采纳率。
0 citationsRead paper
Selected work
Representative Papers
Recent publications
Latest Papers
Chain-of-Thought Faithfulness of Reasoning Models Varies with Where and How Preference Cues Are Delivered
研究通过FACE-Eval评估了不同位置和形式的偏好提示对推理模型忠实度的影响，发现工具返回和隐式提示降低了模型的明确承诺度，增加了未表达采纳率。
0 citationsRead paper