The paper compares Jev with nine language models on the ContractNLI task, assessing inference cost, response time, average correctness, and correctness under repeated requests. Controlled experiments vary hypothesis visibility, requested outputs, and output order while keeping the contract and target judgment fixed. Jev achieves the lowest cost and median response time, whereas hosted language models show higher baseline accuracy, but rankings differ when evaluating correctness across all conditions and repeats.
arXiv:2609.37647v1 Announce Type: cross Abstract: Jev is a commercial System One model from TypeSafe AI that does not generate text: given a state and typed questions, it returns a choice from fixed...
arXiv:2606. 08840v1 Announce Type: new Abstract: Code generation models are typically compared using compact execution benchmarks and aggregate pass rates, but such summaries obscure how performance varies across programming languages, problem families, and failure modes.
The paper introduces LongJudgeBench, a benchmark designed to evaluate large language models (LLMs) acting as judges for long-form text generation. It highlights that long-form evaluation requires complex, document-level assessments beyond simple length, such as organization, coverage, depth, consistency, and scenario-specific quality. Experiments show a significant reliability gap among current LLM judges, indicating instability across scenarios and limited effectiveness of rubrics or references.
arXiv:2608. 20204v1 Announce Type: new Abstract: Legal work, with its heavy reliance on processing large amounts of text, is often considered one of the domains most exposed to the use of LLMs.
JudgeSense is a benchmark comprising 880 items from human‑labelled corpora, each presented under two differently worded instructions that ask the same question. The study evaluates 25 judges from six providers across four tasks, measuring how rewording affects agreement with the judge’s own verdicts. Results show that rewording reduces agreement on all tasks, with significant effects on two, and that stability varies across tasks and is not predicted by parameter count.
The paper reports on building a retrieval‑augmented legal assistant for Uzbek that operates in both a managed cloud service and an on‑premises deployment. It introduces two new domain benchmarks—one for retrieval and one for end‑to‑end QA—and shows that fine‑tuning an open‑weight text embedder (UTE‑1) can close the performance gap with proprietary models under tight cost and latency constraints. The authors also provide negative results for a QLoRA experiment and release the benchmarks, evaluation code, and the fine‑tuned embedder for future low‑resource legal NLP work.
arXiv:2605. 29738v2 Announce Type: replace-cross Abstract: Legal NLP benchmarks overwhelmingly evaluate a single language or aggregate tasks that differ fundamentally across jurisdictions, making cross-lingual comparison impossible.
arXiv:2607. 28282v1 Announce Type: cross Abstract: Evaluating the quality and relevance of textual outputs from Large Language Models (LLMs) remains challenging and resource-intensive.
Legal Research Bench (LRB) is a new benchmark comprising 413 open-ended U.S. legal research questions, each paired with a gold answer, supporting authorities, and a binary grading rubric. The study evaluates thirteen advanced language‑model agents using web search, case‑law search, page parsing, and retrieval tools, scoring responses only when all required criteria are met and cited authorities verify. Results show that even the best model, Claude Opus 4.8, achieves full correctness on only 42.9% of questions, with performance varying by legal area and task complexity, and no clear link between more tool calls or inference cost and higher accuracy.
ContractScrub is a new benchmark that evaluates large language models on the task of contract scrubbing—reviewing legal agreements for errors and inconsistencies. The benchmark includes contracts crafted by experienced lawyers covering diverse error types such as misuse of defined terms, incorrect references, and inconsistent language. Initial tests show that even leading models perform poorly, with only one achieving a 0.75 macro‑average recall, highlighting the gap between general LLM performance and real‑world legal tasks.
The paper presents a deployed system for answering questions over normative documents that is aware of document version and scope. It evaluates a hosted retrieval service against a governed system that applies explicit rules for version and scope resolution, finding the governed system achieves a higher score (97.7 vs 88.1). The study includes a public benchmark, evaluation scripts, and reports commercial deployment metrics, such as 1,126 users and 100,000 calls per day by April 2026.
The paper argues that traditional global calibration metrics, such as Expected Calibration Error and Brier Score, are confounded by differences in model accuracy when comparing large language models. It introduces ACE, an accuracy‑controlled evaluation framework that offers Instance‑Aligned, Distribution‑Aligned, and Candidate‑Aligned views to provide fairer cross‑model comparisons. Experiments across various benchmarks reveal that many reported calibration advantages disappear after accuracy control and that model rankings often reverse, indicating that raw global metrics are unreliable for cross‑model calibration assessment.