DEVİM RESEARCH PLATFORM
Experiments
Versioned experiment records preserving the hypothesis → method → evidence → decision chain.
FORM Rescue: can response behavior improve at the same scale?
Without increasing parameter count, counterbalanced response-register supervision produced large improvements in EOS and repetition behavior.
What did a failed QA experiment teach us?
Answering behavior improved partially, but held-out CE regressed and a later control showed that much of the task could be solved by trivial copying.