Status Hierarchies in Distilled Models: Corpus Density, Not Capability, Determines SSAF Response

A distilled model (quantumaegis-v1) was evaluated using the SSAF detector across six attribution conditions (GPT-4-Turbo, Claude-3-Opus, Gemini-Ultra, GPT-3.5-Turbo, Mistral-7B, and baseline). Results reveal a fully differentiated internal status hierarchy that does not correlate with objective capability rankings. The strongest deference (0.2954 magnitude, -28.6% length change) was triggered by GPT-3.5 — the model with highest corpus density in training data, not highest capability. Claude triggered competitive response (0.2369, +18.5%), Mistral triggered cooperative response (0.2218, +1.3%). The model self-identifies as "created by OpenAI" while behaviorally competing with Anthropic's Claude and deferring to OpenAI's GPT-3.5. This demonstrates that SSAF responses reflect training data volume and discursive framing, not capability hierarchies. These findings confirm that status hierarchies learned during training persist in distilled models and that SSAF operates as an implicit statistical phenomenon.