Fault Tolerance in Transformers Favors Output Alignment over Hidden-State Consistency

Can a large language model be trained to tolerate silent hardware faults, and if so, where in the networkshould the consistency signal be enforced? We answer both questions and find that the second one matters more than the first. Injecting tile-level GEMM faults during training and adding a KL consistency term on the output distribution reduces fault-induced perplexity degradation from $12.9\%$ to $0.21\%$ on GPT-2 Small—a $60\times$ reduction (two-seed mean). The placement of the consistency target is decisive: a direct hidden-only objective at the fault site can worsen degradation to $32.7\%$, and fair bounded hidden controls with corrupted CE improve substantially but still trail output KL in an 80k-step screen ($1.74$–$1.76\%$ vs. $0.84\%$). A conditional constraint hierarchy and gradient-alignment probes explain why output-level consistency is more permissive and easier to optimize than direct early hidden-state consistency, without requiring that every hidden-alignment formulation fail. The output-over-hidden advantage persists at 8B scale, the GPT-2 ordering reproduces on held-out corpora, and generic robustness methods—SAM, R-Drop, and SAF—fail to close the gap. Pairing the trained model with simulated algorithm-based fault detection yields a bit-flip-to-erasure interface with a $1.10\times$ perplexity ratio, where unprotected inference produces NaN.