眼之所惧：以“读盲”写作解释Transformer中的巨大激活
Which the Eye Fears: Writing with Read-Blindness Explains Massive Activations in Transformers
- PsiDagger
- IIT Delhi（德里印度理工学院）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本研究通过算子级机制分析发现，Transformer中巨大激活的持续存在源于注意力和前馈层在读取时忽略MA坐标的“读盲”现象，该现象先于FFN放大出现，且模型主动维持此不对称性。
AI中文摘要：
Transformer中的巨大激活特征（MAs）是残差流中的极值特征，尽管模型有能力抑制它们，它们仍能跨层持续存在。它们为何能存活？我们通过注意力和前馈（FFN）块的算子级机制分析进行的调查揭示，这些块在读取时系统性地忽略MA坐标，但在写入时却不忽略；这造成了一种读写不对称性，阻碍了纠正性反馈，同时允许持续累积。我们发现注意力和前馈层都具有这种读盲性，并共同促成MAs的出现与持续。为验证先前假设FFN的放大能力是MAs主要原因的工作（Sun等人，2026），我们分析了学习过程中的模型检查点。与我们的预期相反，读盲性在FFN放大之前出现，表明它在MA机制中起上游作用。我们进一步贡献了梯度分析，将此行为与损失景观中令人惊讶的不对称性联系起来，得出结论：模型主动维持这种读盲性。最后，我们发现在不同位置移除读阻塞会诱发其他地方的补偿性转移，但MAs仍然持续存在。
英文摘要：
Massive activation features (MAs) in Transformers are extreme-value residual-stream features that persist across layers despite the model's ability to suppress them. Why do they survive? Our investigation using an operator-level mechanistic analysis of attention and feed-forward (FFN) blocks reveals that these blocks systematically ignore MA coordinates while reading, but not while writing; creating a read-write asymmetry that blocks corrective feedback while allowing continued accumulation. We find that both attention and feed-forward layers have this read-blindness, and contribute to the emergence and persistence of MAs. To validate prior work that hypothesized that FFN's amplification abilities is the primary reason for MAs (Sun et al., 2026), we analyze the model checkpoints during learning. Contrary to our expectation, read-blindness emerges before FFN amplification, suggesting that it acts upstream in the MA mechanism. We further contribute gradient analysis to link this behavior to surprising asymmetries in the loss landscape, concluding that the model actively maintains this read-blindness. Finally, we find that removing read-blocking at different locations induces compensatory shifts elsewhere, but MAs still persist.
