Results for “multiple-choice”

3 skills

Benchmarks zero-shot in-context learning on BIG-Bench Hard multiple-choice tasks, comparing self-generated demonstrations against direct prompting and chain-of-thought baselines, and reports accuracy.

More results

Nemo Mbridge Perf Moe Long Context

Provides guidance for training Mixture-of-Experts models with long context windows, covering context parallelism sizing, selective recomputation, dispatcher choices, and practical patterns from recent experiments.

Implements Manifold-Constrained Hyper-Connections (mHC) using Doubly Stochastic Matrices to improve deep learning stability.