dataset
active
dataset:gsm8kGSM8K
Grade school math dataset used for math task in E3.
Neighborhood — ranked by edge-count
Papers (3)
paper
Methods (2)
method
- Quantitative study correlating layer-wise anchoring geometry (S_max, AUS_N) with behavioral thresholds θ50
- Computes reflection direction as mean difference between MLP and attention output representations of first tokens in reflection vs. non-reflection steps