Search papers, labs, and topics across Lattice.
XpertBench is introduced as a new benchmark to evaluate LLMs on complex, open-ended tasks across 80 professional domains, using 1,346 tasks derived from expert submissions. To address evaluation biases, they introduce ShotJudge, an LLM-based evaluation paradigm calibrated with expert few-shot examples. Evaluation of SOTA LLMs on XpertBench reveals a performance ceiling with a peak success rate of only ~66%, highlighting a significant "expert-gap" in current AI systems.
LLMs still fail to demonstrate expert-level proficiency, achieving only ~66% success on a new benchmark of real-world professional tasks spanning finance, healthcare, and law.
As Large Language Models (LLMs) exhibit plateauing performance on conventional benchmarks, a pivotal challenge persists: evaluating their proficiency in complex, open-ended tasks characterizing genuine expert-level cognition. Existing frameworks suffer from narrow domain coverage, reliance on generalist tasks, or self-evaluation biases. To bridge this gap, we present XpertBench, a high-fidelity benchmark engineered to assess LLMs across authentic professional domains. XpertBench consists of 1,346 meticulously curated tasks across 80 categories, spanning finance, healthcare, legal services, education, and dual-track research (STEM and Humanities). These tasks are derived from over 1,000 submissions by domain experts--including researchers from elite institutions and practitioners with extensive clinical or industrial experience--ensuring superior ecological validity. Each task uses detailed rubrics with mostly 15-40 weighted checkpoints to assess professional rigor. To facilitate scalable yet human-aligned assessment, we introduce ShotJudge, a novel evaluation paradigm that employs LLM judges calibrated with expert few-shot exemplars to mitigate self-rewarding biases. Our empirical evaluation of state-of-the-art LLMs reveals a pronounced performance ceiling: even leading models achieve a peak success rate of only ~66%, with a mean score around 55%. Models also exhibit domain-specific divergence, showing non-overlapping strengths in quantitative reasoning versus linguistic synthesis.. These findings underscore a significant"expert-gap"in current AI systems and establish XpertBench as a critical instrument for navigating the transition from general-purpose assistants to specialized professional collaborators.
