Glossary · Term

gradient norm

← all terms

Definition

Plain language

A single number measuring how strong the learning signal is at a given training step.

As stated in the literature

The magnitude of the gradient vector during training; a diagnostic for learning-signal strength, where anomalously small values can reveal silent bugs (as in the DeepSpeed CPU-offload case) or low-variance reward groups.

Also called: gradient norms

Why it matters: It's a quick health check on whether a model is actually learning, helping catch silent failures that would otherwise waste a training run.

For example, if the learning signal suddenly reads near zero when it shouldn't, that tiny gradient norm can tip you off to a hidden bug in the training setup.

Heard on the show

“The practical FPO — the deployable version, the one without the oracle, the "just penalize the gradient norm" one — wins 140 to 135.”
Episode 025 — The Missing Gradient Term That Predicts Sycophancy in RLHF

Mentioned in 3 episodes

  1. 025
    The Missing Gradient Term That Predicts Sycophancy in RLHF
  2. 010
    When Reward Climbs But Reasoning Goes Generic: Diagnosing Template Collapse in Agentic RL
  3. 009
    How Two Silent Library Bugs Quietly Invalidated a Wave of Reasoning Papers

Related terms