Author

Anish Athalye

2 papers indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

#artificial intelligence Apr 2026

An Imperfect Verifier is Good Enough: Learning with Noisy Rewards

The results indicate that imperfect verification can support effective RLVR in the tested settings, while aggregate error rates alone do not characterize the learning signal.

Andreas Lindhardt Plesner, F. Guzmán, Anish Athalye · arXiv.org · 10 citations · ⚡1

#artificial intelligence Preprint Sep 2026

A Cheap Verifier is Good Enough: LLM Post-training is Robust to Erroneous Rewards

This paper compares two low-cost choices retrospectively -- a cost-reducing choice and a balanced choice -- with estimated grading cost reductions and average post-training score gaps of 1-3 points from the best evaluated training verifier, and establishes that verifier choices are interchangeable.

Andreas Lindhardt Plesner, Curtis G. Northcutt, F. Guzmán et al. · 0 citations