We measure social cognition in language models by testing their ability to select socially appropriate responses in controlled multiple-choice scenarios. Across 50 position-randomized vignettes evaluated on nine model-condition pairs spanning 7B to 72B parameters, we find that RLHF instruction tuning hurts social cognition at 7B (-18 percentage points, from 72% to 54%) but helps at 72B (+6 points, from 84% to 90%). The deficit is not one of knowledge but of processing mode: instruction-tuned models at small scale optimize for literal compliance at the expense of the fuzzy contextual pattern matching that social reasoning requires. Forcing models to explain their reasoning costs an additional ~8 points at both scales — a rationalization bias in which explicit deliberation overrides correct intuitive judgment. Stochastic resonance probe confirms the social signal is present but suppressed in instruction-tuned models. The result reframes RLHF alignment not as a capability enhancement but as a noise-tolerance tradeoff: compliance training reduces the variance that social cognition depends on.

Paper 23 in the Structural Compression Theory research program. Series: Activation Geometry.