concept
active
concept:update-and-forget-loopUpdate-and-forget loop
Small models learn then lose rare-task updates between observations.
Neighborhood — ranked by edge-count
Papers (1)
paper
Concepts (1)
concept
- Gradient interferenceextendsFrequent-task updates overwrite rare-task features.
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.
- Key dimension analyzing how wide the gulfs of execution and evaluation are in a system and how they relate; uses concepts from The Design of Everyday Things.
- Second stage of DiffLogic CA where a DLGN computes each cell's new binary state from perception output and current state
- Repetitive behavioral pattern observed under high steering strengths in SAE feature self-steering experiments
- Update strategy where random subsets of cells are updated per step, simulating independent cell clocks
- Default update strategy where all cells update simultaneously; contrasted with asynchronous updates
- Training regime where random subsets of cells update per step, improving robustness of learned circuits
- A dynamic programming method for computing optimal value functions and policies in known MDPs.
- Recursive inference of hidden states and parameters using free energy minimization.