method
active
method:screen-stage-unexpectedness-ratingScreen stage (unexpectedness rating)
Step 2 flags unexpected behaviors with a 1-5 rubric and frozen classifier question.
Neighborhood — ranked by edge-count
Papers (1)
paper
Methods (1)
method
- Agentic pipeline that discovers unexpected model behaviors and investigates them with counterfactual prompt edits.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- First stage of DiffLogic CA update where each cell gathers information from neighboring cells via logic gate kernels
- Step 3 runs 5-15 counterfactual experiments.
- Step 4 judge scores support on 1-10 scale.
- Step 1 of CHIVE.
- Stage 2 of character training: DPO from teacher model to student model to transfer desired behavioral expressions
- Stage 3 of character training: SFT on synthetic introspective data generated by post-distillation checkpoint
- Second stage of DiffLogic CA where a DLGN computes each cell's new binary state from perception output and current state
- Aversive stimuli are defined as surprising, linking punishment to prediction failure.