Codifying the Judge: Scalable Evaluation via Program Distillation
This paper introduces program distillation as a simple, efficient alternative to LLM-as-a-judge, distilling decision logic into a committee of programs for transparent, low-cost evaluation. The PAJAMA system aggregates program verdicts and falls back to an LLM when confidence is low, matching the performance of a 13B-size LLM judge across five datasets and four model families. On RewardBench, a reward model distilled from program verdicts outperforms one trained on proprietary LLM labels at two orders of magnitude lower API cost.
-->
[Submitted on 29 May 2026]
Title:Codifying the Judge: Scalable Evaluation via Program Distillation
View a PDF of the paper titled Codifying the Judge: Scalable Evaluation via Program Distillation, by Tzu-Heng Huang and Shengqi Qiu and Frederic Sala
View PDF HTML (experimental)
Abstract:LLM-as-a-judge has become the standard for automated evaluation, but it suffers from high cost, significant latency, and opaque decisions -- limitations that undermine its scalability and reliability. We address these with a simple, efficient alternative: program distillation. Instead of prompting an LLM at the evaluation time, we distill its decision logic into a committee of programs that score candidates directly. These programmatic judges offer transparency, are easily inspected or edited, and eliminate per-sample API costs. Building on this notion, we introduce PAJAMA, a system that synthesizes programs as judges, aggregates their decisions into a joint verdict, and incorporates a fallback mechanism to selectively escalate low-confidence cases to an LLM. Across five datasets and four model families, we show that programmatic judges can match the performance of a 13B-size LLM judge. When using program outputs as routing signals, PAJAMA improves both accuracy and throughput and advances the Pareto frontier. Beyond evaluation, programmatic judges produce cheap and effective reward signals: on RewardBench, a reward model distilled from programs' verdicts outperforms one trained on a proprietary LLM's labels at two orders of magnitude lower API cost.
Subjects:
Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as: arXiv:2607.22561 [cs.AI]
(or arXiv:2607.22561v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2607.22561
arXiv-issued DOI via DataCite
Submission history
From: Tzu-Heng Huang [view email] [v1] Fri, 29 May 2026 03:13:24 UTC (9,073 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled Codifying the Judge: Scalable Evaluation via Program Distillation, by Tzu-Heng Huang and Shengqi Qiu and Frederic Sala
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.AI
new | recent | 2026-07
Change to browse by:
cs cs.LG
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
Loading...
Data provided by:
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)