Paper Detail

Themis: Training Robust Multilingual Code Reward Models for Flexible Multi-Criteria Scoring

Indraneil Paul, Glavaš Glavas, Iryna Gurevych

huggingface Score 11.4

Published 2026-05-01 · First seen 2026-05-04

General AI

Abstract

Reward models (RMs) have become an indispensable fixture of the language model (LM) post-training playbook, enabling policy alignment and test-time scaling. Research on the application of RMs in code generation, however, has been comparatively sparse, with existing work largely focusing on execution feedback. This choice constrains post-training to optimizing functional correctness over self-contained executable code. In this work, we examine the training and evaluation of multilingual, multi-criteria code RMs. To this end, we first compile Themis-CodeRewardBench, a benchmark to evaluate code RMs across five preference dimensions (i.e., criteria) and eight programming languages, on which we profile 50+ code, math, and general-purpose RMs. Observing the limited proficiency of current RMs beyond scoring for functional correctness, we develop Themis-CodePreference, the largest open-source collection of code preferences to date (more than 350k preference pairs), and use it to train Themis-RM, a suite of multilingual code reward models for flexible multi-criteria scoring, ranging in size from 600M to 32B parameters. Our experiments and ablations demonstrate positive scaling trends, strong cross-lingual transfer when training on diverse preferences, and the importance of multi-criteria training for reliable code reward modeling.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
now
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@misc{paul2026themis,
  title = {Themis: Training Robust Multilingual Code Reward Models for Flexible Multi-Criteria Scoring},
  author = {Indraneil Paul and Glavaš Glavas and Iryna Gurevych},
  year = {2026},
  abstract = {Reward models (RMs) have become an indispensable fixture of the language model (LM) post-training playbook, enabling policy alignment and test-time scaling. Research on the application of RMs in code generation, however, has been comparatively sparse, with existing work largely focusing on execution feedback. This choice constrains post-training to optimizing functional correctness over self-contained executable code. In this work, we examine the training and evaluation of multilingual, multi-cr},
  url = {https://huggingface.co/papers/2605.00754},
  keywords = {reward models, language models, code generation, functional correctness, multilingual, multi-criteria, preference dimensions, cross-lingual transfer, parameter-efficient fine-tuning, code available, huggingface daily},
  eprint = {2605.00754},
  archiveprefix = {arXiv},
}

Metadata

{}