语言模型与大脑对齐吗？预测分数并不足够
Do Language Models Align with Brains? Prediction Scores Are Not Enough
AI总结：
本文通过L-PACT框架评估语言模型与大脑的对齐性，发现预测分数不足以支持两者对齐的结论，所有测试结果均被控制条件解释。
摘要：
大脑-语言模型比较常将神经预测分数视为模型表征捕捉大脑相关语言计算的证据。我们探讨了语言模型是否与大脑对齐，以及预测分数是否足以支持这一主张，使用L-PACT框架评估预测、关系、机制剥离和可靠性限制的证据。在自然语言神经数据集和衍生语言模型表征上，L-PACT比较了真实模型特征与干扰基线和严苛对照，测试模型-大脑轮廓是否重现大脑-大脑模式，重新计算机制剥离后的测试分数，并将证据归一化以对抗大脑-大脑上限。锁定分析集包含414个预测控制行、2304个关系轮廓行、4320个机制剥离行、420个大脑-大脑上限行和146个整合决策行。实验敏感性检查显示，大脑-大脑可靠性、大脑作为模型的运行间关系轮廓、独立低级神经和WAV衍生的声学包络门控，以及确定性植入信号模拟可以产生阳性证据。然而，没有真实模型行通过预测、关系、机制剥离或操作性图灵限制可靠性门；所有146个整合行都被控制解释。更宽松的单一标准规则会计算原始阳性预测、关系、剥离delta和上限归一化效应，但L-PACT将其降级，因为控制解释了明显证据。在分析的衍生人工集上，测试的语言模型表征不满足L-PACT对齐门；明显阳性被转换为可审计的控制解释分类，而不是视为结构对齐。
英文摘要：
Brain-language model comparisons often interpret neural prediction scores as evidence that model representations capture brain-relevant language computation. We asked whether language models align with brains, and whether prediction scores are enough to support that claim, using L-PACT, a source-audited framework that evaluates predictive, relational, mechanism-stripping, and reliability-bounded evidence. Across primary naturalistic language neural datasets and derived language-model representations, L-PACT compared real model features with nuisance baselines and severe controls, tested whether model-to-brain profiles reproduced brain-to-brain patterns, recomputed held-out scores after mechanism stripping, and normalized evidence against brain-brain ceilings. The locked analysis set contains 414 predictive-control rows, 2304 relational profile rows, 4320 mechanism-stripping rows, 420 brain-brain ceiling rows, and 146 integrated decision rows. Assay-sensitivity checks showed that brain-brain reliability, brain-as-model run-to-run relational profiles, independent low-level neural and WAV-derived acoustic-envelope gates, and a deterministic implanted-signal simulation can produce positive evidence when expected. Nevertheless, no real model row passed the predictive, relational, mechanism-stripping, or operational Turing-bounded reliability gates; all 146 integrated rows were control-explained. Less stringent single-criterion rules would have counted raw positive predictive, relational, stripping-delta, and ceiling-normalized effects, but L-PACT downgraded them because controls explained the apparent evidence. In the analyzed derived artifact set, the tested language-model representations do not satisfy L-PACT alignment gates; apparent positives are converted into an auditable control-explained taxonomy rather than treated as structural alignment.
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。
