Evaluating RAG with LLM as a Judge

Expert Support case study: Bolstering a RAG app with LLM-as-a-Judge

Judge Arena: Benchmarking LLMs as Evaluators

When the Judge Changes, So Does the Measurement: Auditing LLM-as-Judge Reliability

arXiv:2607. 08535v1 Announce Type: cross Abstract: An LLM-as-judge score can move even when the candidate responses stay fixed, simply because the evaluator has changed.

The First Token Is Not the Verdict: Hidden Costs of Reading LLM Judges Without Generating

The paper demonstrates that reading a large language model (LLM) judge’s verdict from the logits of its first generated token—an approach used in constrained decoding and likelihood‑scoring evaluation—introduces a significant distortion in position bias. Because judges do not always start with a verdict token (12–49% of cases for Qwen3 judges and <3% for Llama‑3.1‑8B and Phi‑3.5‑mini), this readout often returns the first response rather than a true judgment, inflating position bias by up to 42 points while barely affecting judge accuracy. The authors recommend reporting the frequency with which a judge leads with a verdict token to provide a more accurate assessment of position bias.

Diagnosing LLM Arbitration Behavior over Pre-evidence Epistemic States in RAG-based Fact-Checking

arXiv:2606. 01120v1 Announce Type: new Abstract: In RAG-based fact-checking, LLMs are increasingly used as verifiers to check given claims against retrieved evidence.

JudgeMoE: Distributional Aggregation for LLM-as-a-Judge

arXiv:2610.07109v1 Announce Type: new Abstract: When an LLM judge scores an output, its score distribution retains uncertainty and disagreement information that is lost after scalar compression. We i...

LLM Judge Validation Under Sparse Overlap: From Inference to Design

arXiv:2609.31857v2 Announce Type: replace Abstract: Validating an LLM-as-a-judge requires estimating its agreement with humans, yet annotation budgets rarely allow every item to be multiply labeled....

LLM Judges Have Dark Current: A Psychometric Datasheet for LLM-as-a-Judge Evaluation

arXiv:2606. 15610v1 Announce Type: cross Abstract: LLM-as-a-judge systems are now routinely used for open-ended model evaluation, where human preference annotation is costly, slow, and difficult to reproduce.

Three Ways Classical Test Theory Can Mislead About LLM Judges

The article examines how classical test theory statistics—Kuder‑Richardson coefficient, dependability index, and Livingston‑Lewis accuracy—can mislead when applied to large language model (LLM) judges that are evaluated with a single prompt and no gold labels. Using Claude Haiku 4.5 on 210 short‑answer items, the authors show that these metrics fail to isolate the judge’s performance because the judge’s single administration provides no variance component. They argue that reliable statements about an LLM judge require gold labels or varied scorer facets, and that bank design heavily influences reliability estimates.

Auditing and Repairing LLM-as-Judge Failures in a Production Text-to-SQL Pipeline

The paper evaluates a production text‑to‑SQL pipeline that uses an LLM as a judge, finding that the deployed gpt‑4o‑mini judge agrees with human annotators only weakly (Cohen’s kappa 0.04 on a disagreement‑enriched set and 0.42 on a random spot‑check). The authors identify a specific failure mode, GRADE‑HALLUCINATION, responsible for most over‑flags, and demonstrate that a self‑hosted Qwen3.6‑27B model achieves substantially higher agreement (kappa 0.72) at a much lower cost. They also show that ensembling judges does not improve performance, and that their audit method flags a significant portion of out‑of‑domain SQLs as potential issues.

A Judge Should Know What Changed:Construct Validity for LLM-as-a-Judge Evaluation

The paper introduces a two‑dimensional construct validity framework for evaluating large language models (LLMs) as judges, defining invariance (S) and sensitivity (R) to construct‑preserving and construct‑changing edits. Experiments across seven judges and four domains reveal high invariance (average S = 0.945) but low sensitivity (average R = 0.319), with sensitivity varying by edit type. Audits of public label sets show that surface‑only predictors can reproduce a substantial portion of labels, underscoring that high agreement does not guarantee construct validity.

MAWILE: Multi-Axis Workbench for Inspecting LLM Evaluators

arXiv:2609.22599v1 Announce Type: new Abstract: Large language model (LLM) judges provide a flexible and scalable method for evaluating model and agent outputs, but their verdicts can be sensitive to...