从可解释性工件中学习自我解释：在向量-标签对上训练轻量级适配器
Learning Self-Interpretation from Interpretability Artifacts: Training Lightweight Adapters on Vector-Label Pairs
AI总结：
通过训练轻量级适配器（标量仿射适配器，仅需d_model+1参数）在可解释性工件上，保持语言模型完全冻结，实现了跨任务和模型族的可靠自我解释，在稀疏自编码器特征标注、主题识别和多跳推理桥接实体解码等任务上显著优于未训练基线。
摘要：
自我解释方法促使语言模型描述其内部状态，但由于超参数敏感性而仍然不可靠。我们表明，在可解释性工件上训练轻量级适配器，同时保持语言模型完全冻结，可以在任务和模型族中产生可靠的自我解释。一个仅需$d_\text{model}+1$个参数的标量仿射适配器就足够了：训练后的适配器生成稀疏自编码器特征标签，其性能优于训练标签本身（在70B规模下，生成评分为70% vs 50%），以94%的召回率@1识别主题（未训练基线为1%），并在多跳推理中解码既不在提示中也不在响应中出现的桥接实体，从而无需思维链即可揭示隐式推理。仅学习到的偏置向量就占了改进的85%，更简单的适配器比更具表达力的替代方案具有更好的泛化能力。通过提示描述控制模型知识，我们发现从7B到72B参数，自我解释的提升超过了能力提升。我们的结果表明，自我解释随着规模扩大而改善，且无需修改被解释的模型。
英文摘要：
Self-interpretation methods prompt language models to describe their own internal states, but remain unreliable due to hyperparameter sensitivity. We show that training lightweight adapters on interpretability artifacts, while keeping the LM entirely frozen, yields reliable self-interpretation across tasks and model families. A scalar affine adapter with just $d_\text{model}+1$ parameters suffices: trained adapters generate sparse autoencoder feature labels that outperform the training labels themselves (70% vs 50% generation scoring at 70B scale), identify topics with 94% recall@1 versus 1% for untrained baselines, and decode bridge entities in multi-hop reasoning that appear in neither prompt nor response, surfacing implicit reasoning without chain-of-thought. The learned bias vector alone accounts for 85% of improvement, and simpler adapters generalize better than more expressive alternatives. Controlling for model knowledge via prompted descriptions, we find self-interpretation gains outpace capability gains from 7B to 72B parameters. Our results demonstrate that self-interpretation improves with scale, without modifying the model being interpreted.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
