Tag: rnn
-
Portfolio ·
The No-Deal Button — Training-Environment Echoes in Trained Agents
Five phases and a 50-combination sweep (27,500 rounds) of Alice & Bob self-play. The word 'button' surfaces 95 times across the sweep — a UI mechanism from the training environment that doesn't exist at inference. Temperature-gated, category-stable, concept-persistent under entropy. Announced-but-broken tools still reshape behaviour. The training environment leaves imprints in the model's navigable state space.
-
Portfolio ·
The No-Deal Button — Log Archaeology
Phase-0 forensic work for the No-Deal Button study. Traces Alice & Bob agents' reference to a 'no-deal button' back to 316 instances of the word in Meta's Deal-or-No-Deal training corpus — including a misspelled variant 'buttin' that predicts the deformation pattern the agents themselves produce at high sampling temperature.