Skip to content
@ishida-lab

ishida lab @ UTokyo

We are a sub-group of Machine Learning and Statistical Data Analysis Lab (mslab) in UTokyo, with a focus on model evaluation, alignment, and safety.

Popular repositories Loading

  1. irreducible irreducible Public

    [ICLR 2023] Is the Performance of My Deep Network Too Good to Be True? A Direct Approach to Estimating the Bayes Error in Binary Classification

    Python 25

  2. capbencher capbencher Public

    [ICML 2026] CapBencher toolkit: Give your LLM benchmark a built-in alarm for leakage and gaming

    Python 12 1

  3. ProbabilityBench ProbabilityBench Public

    ProbabilityBench: Evaluating Jev and Other Decision Models on True Probabilities

    Python 11 1

  4. iw-dpo iw-dpo Public

    [TMLR 2025] Importance Weighting for Aligning Language Models under Deployment Distribution Shift

    Python 5

  5. capcode capcode Public

    [NeurIPS 2026] CapCode: Detecting cheating in coding agents with capped, randomized tests

    Python 5

  6. capreward capreward Public

    [NeurIPS 2026] CapReward: Penalizing implausibly high pass rates to prevent reward hacking in coding RL

    Python 5

Repositories

Showing 6 of 6 repositories

Top languages

Loading…

Most used topics

Loading…