verifier

Definition

A separate model or program that grades whether an answer is correct.

A model or deterministic checker that scores candidate outputs, used to provide reward signals or filter rollouts in training and evaluation.

Also called: verifiers

Mentioned in 14 episodes

1. 078
   Training a Markdown File: When LLM Self-Improvement Borrows the Discipline of Neural Net Training
2. 076
   Same Model, Organized Differently: How an Agent Architecture Beat Frontier Systems at Research Math
3. 075
   Growing Code and Proof Together: Verified Systems in Ten Hours Instead of a Year
4. 071
   When the Model Is Fine and the Plumbing Is Broken: Fixing Agents at the Interface
5. 067
   An AI Just Solved a 1996 Erdős Problem—and the Simplest Agent Won
6. 062
   Treating Hallucinations as Exploits: A Gate-Based Architecture for Agent Safety
7. 060
   When Splitting One Model Across Three Agents Doubles Its Accuracy
8. 044
   How One Sentence and a Forged History Flip the Most Aligned Models
9. 028
   Teaching a Model to Hire Copies of Itself: Recursive Agent Optimization
10. 027
    When AI Agents Build the Serving Stack: A Bet on Bespoke Infrastructure
11. 024
    An AI Agent That Found 28 Zero-Days in Windows — And What Made It Work
12. 019
    When the Best Reward Model Trains the Worst Policy: Inside EvoLM
13. 017
    When the Agent Grades Its Own Homework: A Brutal New Benchmark for AI Workers
14. 011
    When RL Actually Teaches Agents Something New, And When It Doesn't

Related concepts