Skip to content

Author

Valmik Nahata

We have 2 of 4 papers

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 Preprint Oct 2026

Simple Features Beat Most AI Control Monitors

It is shown that on a standard evaluation dataset, simple features beat most monitors without ever running the code, outranking 20 of 24 open-weight models prompted as monitors, suggesting that a high score on such a dataset is not by itself evidence of sabotage detection.

Shubhrangshu Debsarkar, Valmik Nahata · 0 citations

Last Translation Benchmark

The Last Translation Benchmark is introduced, a collection of human-authored and peer-reviewed examples that break leading machine translation models and a new evaluation approach: each example comes with handcrafted verification rules describing concrete failure cases on that example, therefore allowing reliable and a...

Vilém Zouhar, Niyati Bafna, Mukund Choudhary et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.