Human-in-the-Loop Evaluation

Follow this topic Ask about this topic

This field examines how people evaluate AI outputs and collaborate with AI during testing and deployment, including whether systems are accurate, useful, safe, and appropriately trusted. Common approaches include expert rubrics and pairwise preferences, user studies and interaction traces, human-validated benchmarks, and LLM judges calibrated against human ratings, with uncertain or disputed cases escalated for review. Researchers also study how to allocate limited review effort, model annotator disagreement, and measure effects on human decisions and learning over time. The central tension is scaling evaluation with automated judges and simulations without losing expert context or mistaking agreement, confidence, or benchmark scores for real-world quality.

Latest papers

160 papers19 papers in the last four weeks, level with the four weeks before. 0.1% of all new papers.

HALO: Human-Grounded Granular Evaluation of Proactive Agents
Oct 29, 2026Wenqi Pei, Xinyu Wang, Jinrui Liu +5Proactive Agents Proactive Assistance
University of California, Davis · 2077AI · University of Southern California
Human Evaluation with Automatic Scores: Separating Sampling and Estimation Gains
Oct 28, 2026Yang Liu, Ran Zou, Zhengtao Gui +1Human-in-the-Loop Evaluation LLM Evaluation
Department of Statistics and Operations Research, University of North Carolina at Chapel Hill, USA · Department of Statistics, University of California, Irvine, USA · School of Computer Science and Engineering, University of New South Wales, Australia
From Self-Confidence to Statistical Accuracy: Scaling Human-in-the-Loop Evaluation with Direct and LLM-Assisted Auditing
Oct 26, 2026Harine Choi, Chenchen Kuai, Zihao Li +5LLM Evaluation LLM Auditing
Texas A&M Transportation Institute, Texas A&M University System · Zachry Department of Civil and Environmental Engineering, Texas A&M University · Department of Civil, Construction and Environmental Engineering, Marquette University +2
DealsFlow: An Evidence-Gated Agentic Harness for Venture Capital Deal Triage
Oct 23, 2026Ming Ze Tang, David Cutting, Boyd Carson +1Financial Language Model Evaluation LLM Agent Evaluation
Queen’s University Belfast, Belfast, United Kingdom · Sapphire Capital Partners, Belfast, United Kingdom
SciDiagBench: A Type-Specific Benchmark for Human-Aligned Evaluation of Scientific Diagrams
Oct 20, 2026Hao Wang, Zixuan Li, Long Bai +4VLM Evaluation Image Generation Evaluation
Institute of Computing Technology, Chinese Academy of Sciences · State Key Laboratory of AI Safety · School of Advanced Interdisciplinary Sciences, University of Chinese Academy of Sciences +2
ValiAssist: Scaling LLM-generated Programming Feedback Validation by Demonstration
Oct 19, 2026Xiaohang Tang, Rehema Abulikemu, Peng Gao +1Generative AI Evaluation Human-in-the-Loop Evaluation
Virginia Tech, USA
Selecting the Best Few with Human Evaluators: Adaptively Combining Ratings and Pairwise Comparisons
Oct 19, 2026Motti Goldberger, Nils RudiHuman-in-the-Loop Evaluation Preference Elicitation
Yale School of Management
VEILED-SH: Vulnerability Expression through Indirect Language Evaluation Dataset (Self-Harm)
Oct 19, 2026Leanne Tan, Keith Chua, Shaun Khoo +7LLM Safety Benchmarks Language Model Safety Evaluation
GovTech Singapore · MOH Office for Healthcare Transformation · Singapore University of Technology and Design +2
Towards Human-AI Complementarity for AI Safety
Oct 18, 2026Sander Lanbo Shi, Miranda Anna Christ, Eric Modesitt +20Human-AI Complementarity Human Oversight of AI
Sampura Research · University of California, Berkeley · University of Illinois Urbana-Champaign +6
Edit Discovery in Human-AI Interactions
Oct 15, 2026Alireza Mirrokni, Shayan Kiyani, George Pappas +1Human-in-the-Loop Evaluation Human-AI Interaction
EPFL · University of Pennsylvania
Author-in-the-loop Agentic Reproducibility Auditing for a Machine Learning Conference
Oct 15, 2026Yukun Dong, Linjun Zhang, Tian ZhengML Reproducibility Scientific Reproducibility
Rutgers University · Columbia University
Kolab: Benchmarking Human-Agent Collaboration in Long-Horizon Workflows
Oct 15, 2026Shirley Anugrah Hayati, Patricia Anugrah Setiani, Ruizi Wang +4Human-AI Collaboration Long-Horizon Agent Evaluation
Department of Computer Science, University of Minnesota · Department of Economics, University of Minnesota
From Human Corrections to Learned Segmentations: A Critical Mathematical Review of Information, Learning, and Human Effort
Oct 14, 2026Jon SporringHuman-in-the-Loop AI Human-in-the-Loop Annotation
Department of Computer Science, University of Copenhagen, DK-2100 Copenhagen, Denmark
One-Sided Verification Under Limited Expert Review
Oct 13, 2026Fan Nie, Haotian Ye, Shurui Liu +3Reasoning Verification Human-in-the-Loop Evaluation
Stanford University
Estimating Uncoded Crash Factors with Tabular Foundation and System One Models: Kumo Tabular and Jev
Oct 7, 2026Amir Rafe, Subasish DasRare-Event Estimation Tabular Foundation Models
Texas State University, San Marcos, USA
A Scoping Review and Experimental Study on Reinforcement Learning from Human Feedback for Human-Robot Collaboration
Oct 7, 2026Alexandra Coroiu, Andrea Vogt, Viktor Werbilo +3RL from Human Feedback Human-Robot Collaboration
DLR Institute for AI Safety and Security, Wilhelm-Runge-Straße 10, 89081, Ulm, Germany · DLR Institute for AI Safety and Security, Rathausallee 12, 53757, Sankt Augustin, Germany
Evaluating human-AI workflows for field research in viticulture
Oct 6, 2026Niko Carvajal Janke, Daoyuan Jin, Shivranjani Baruah +4AI-Assisted Scientific Research Human-in-the-Loop Evaluation
Plant Pathology and Plant-Microbe Biology Section, School of Integrative Plant Science, Cornell AgriTech, Cornell University, Geneva, New York, 14456, USA · School of Electrical and Computer Engineering, Cornell University, Ithaca, New York, 14850, USA · Horticulture Section, School of Integrative Plant Science, Cornell AgriTech, Cornell University, Geneva, New York, 14456, USA +1
What Did the Agent Actually Do? Evidence-Grounded Oversight for Long-Horizon Agents
Oct 5, 2026Zhongxiang Sun, Jiahao Yan, Hongkang Zhao +5Software Engineering Agents Human-in-the-Loop Evaluation
Gaoling School of Artificial Intelligence, Renmin University of China · Kuaishou Technology
Designing the Future of User Feedback for Generative AI
Oct 2, 2026Alisa Frik, Julia Bernd, Amitis Karami +1Generative AI Evaluation Human-in-the-Loop Evaluation
International Computer Science Institute, Berkeley, CA, USA · University of California, Berkeley, Berkeley, CA, USA · eBay Inc., San Jose, CA, USA
Auditing Web Agent Evaluation on WebArena-Lite: Human Review of Outcomes and Trajectories
Oct 1, 2026Chengguang Gan, Zimeng He, Yoshihiro Tsujii +3AI Agent Auditing Web Agent Benchmarks
Techtouch, Inc. · Institute of Science Tokyo
Ontology-Based Contextual AI Evaluations (OB-CAIE) Methodology
Sep 30, 2026Julie Krugler Hollek, Michael Zargham, Mala KumarHuman-in-the-Loop Evaluation Automated Evaluation
Humane Intelligence · Dynamical Systems Group
JuryFlow: Disagreement-Guided Human-in-the-Loop Multi-Agent Evaluation
Sep 30, 2026Mufeng Yang, Junwei Yu, Yepeng DingLLM-as-a-Judge Human-in-the-Loop Evaluation
The University of Tokyo Tokyo, Japan · Hiroshima University Higashihiroshima, Japan
Designing Reliable LLM-as-a-Judge Measurement Systems for Multi-Turn Business Agents
Sep 27, 2026Kaiwen Luo, Ming GaoLLM-as-a-Judge LLM Agent Evaluation
Meta
PANEL: An Open-Source, Self-Hosted Web Platform for Human Evaluation of Generative Models
Sep 25, 2026Matteo Spanio, Andrea Poltronieri, Martín RocamoraHuman-in-the-Loop Evaluation Generative AI Evaluation
University of Padova, Padova, Italy · Music Technology Group, Universitat Pompeu Fabra, Barcelona, Spain
Robots That Take Initiative: A Framework for Building and Evaluating Proactive Robots
Sep 24, 2026Maithili Patel, Sonia ChernovaHuman-in-the-Loop Evaluation Robotic Policy Evaluation
School of Interactive Computing, Georgia Institute of Technology, USA
Human-AI-Powered Hypothesis Testing: Cost-Aware Selective AI Scoring and Sequential Human Escalation
Sep 24, 2026Dae Woong Ham, Xuejun Zhao, Stefanus Jasin +1Cost-Aware Inference Human-in-the-Loop Evaluation
Spatial Action Review: A Visual Analytics Dashboard for Auditing Language-to-Action Hand-offs in Electron Microscopy
Sep 21, 2026Samia Mohinta, Albert CardonaHuman-in-the-Loop Evaluation Multimodal QA
University of Cambridge, UK · MRC Laboratory of Molecular Biology, UK
Toward Human-in-the-Loop Robot Failure Recovery: Bridging Communication Gaps in Human-Robot Collaboration
Sep 21, 2026Promise Ekpo, Teju Vijay, Dhruv Mandalik +5Human-in-the-Loop Evaluation Robot Failure Recovery
Cornell University, New York, New York, USA · Cornell Tech, New York, New York, USA · Brown University, Providence, Rhode Island, USA
EnterpriseVal: Quantifying the Efficacy, Reliability and Value of Generative AI in the Enterprise
Sep 18, 2026Abbas Raza Ali, Muhammad Ajmal Siddiqui, Moona ZahidHuman-in-the-Loop Evaluation LLM Reliability
Citigrounp, Inc, London, United Kingdom · Ernst & Young LLP, London, United Kingdom · NVIDIA Corporation, London, United Kingdom
Clinician-Grounded Quality Assurance for AI-Assisted Psychiatric Intake
Sep 17, 2026King Shi, Amanda Li, Jonathan Ivey +7Clinical Language Model Evaluation Mental Health
Johns Hopkins University

Oct 29, 2026J/K move · Enter open · S save

HALO: Human-Grounded Granular Evaluation of Proactive Agents

Wenqi Pei, Xinyu Wang, Jinrui Liu, Cynthia Zhuge+4

University of California, Davis · 2077AI · University of Southern California

Proactive Agents Proactive Assistance Human-in-the-Loop Evaluation

Open paper AskPDF arXiv

Proactive agents must decide whether, when, and how far to assist without an explicit request. Existing evaluations often rely on automated scores or binary verdicts, offering limited human diagnosis of evidence use and intervention quality. To address these gaps, we introduce HALO, Human-Grounded GrAnular EvaLuation of PrOactive Agents. HALO-Arena covers 4 proactive need types, Dependency, Blockage, Risk, and Opportunity, across 10 everyday domains and 50 personal environments. Scenarios require agents to retrieve evidence from structured databases or unstructured documents before judging whether assistance is warranted and choosing its scope. On HALO-Arena, we construct HALO-8K, a dataset of 8,000 responses from 10 models across 400 scenarios, annotated by 20 experts for action level, acceptability verdicts, and human diagnosis. For rejected responses, human diagnosis covers timing, intensity, and other issues. Recorded reasoning tokens and retrieval rounds measure the effort behind these responses. The best proprietary and open-source models achieve 72.88% and 56.63% acceptability, respectively. Our central finding is that over-proactivity is the dominant observed failure mode, with Premature timing or Excessive intensity accounting for 61.20% of rejected responses. Further analyses yield complementary findings on the balance between intervention and restraint, evidence acquisition, and the calibration of timing and intensity.