Commit History
-
Dedupe echoed user text in predictions
Verified signatureb132b97 -
Map run id to deployed checkpoint and add reward distributions
Verified signaturec6c6f2b -
Add grouped result plots to technical report
Verified signatureac245dc -
Fix checkpoint id for Interactive-RL default
Verified signaturedfb59d7 -
Use final LoRA checkpoint in RL default
Verified signaturee4e0cb9 -
Initial poor mans interaction models demo
Verified signature128f88c