concept
active
concept:nucleus-samplingNucleus Sampling
Decoding strategy used throughout experiments; p=0.9 selected to increase response diversity
Neighborhood — ranked by edge-count
Papers (2)
paper
Methods (1)
method
- Truncate CoT and see whether the answer changes.
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.
- New semantic diversity dimension added to prior finding that nucleus sampling is more lexically diverse
- Procedure for sampling 64 random nonnegative combinations of cone basis vectors to evaluate the full cone distribution
- The mechanism by which LLMs generate text: drawing a token from the next-token distribution and appending it to context repeatedly
- A technique to filter model outputs; Redwood Research's project mentioned.
- A Bayesian exploration strategy that samples from the posterior distribution over model parameters to decide actions.
- Dividing feature activation spectrum into 11 evenly-spaced intervals and sampling uniformly to evaluate monosemanticity across activation levels
- Counts c_t ~ Multinomial(S,o_t).
- Baseline using Binomial(k,p) draws from precomputed rollouts.