LAUNCH Lab

university

https://launch.eecs.umich.edu/

AI & ML interests

Factuality, reasoning, alignment, LLM applications

Recent Activity

mkhalifa updated a dataset about 20 hours ago

launch/thinkprm-1K-verification-cots

yujianll submitted a paper 12 days ago

How Well Do Agentic Skills Work in the Wild: Benchmarking LLM Skill Usage in Realistic Settings

jpeper published a dataset about 2 months ago

launch/LudoBench

View all activity

Papers

Gaming the Judge: Unfaithful Chain-of-Thought Can Undermine Agent Evaluation

View all Papers

updated a dataset about 20 hours ago

launch/thinkprm-1K-verification-cots

Viewer • Updated about 20 hours ago • 1k • 40 • 7

submitted a paper to Daily Papers 12 days ago

How Well Do Agentic Skills Work in the Wild: Benchmarking LLM Skill Usage in Realistic Settings

Paper • 2604.04323 • Published 14 days ago • 40

published a dataset about 2 months ago

launch/LudoBench

Viewer • Updated Mar 1 • 638 • 278

published a Space about 2 months ago

LudoBench

Multimodal Game Reasoning Benchmark [ICLR 2026]

updated a dataset about 2 months ago

launch/LudoBench

Viewer • Updated Mar 1 • 638 • 278

updated a Space about 2 months ago

LudoBench

Multimodal Game Reasoning Benchmark [ICLR 2026]

submitted a paper to Daily Papers 3 months ago

Gaming the Judge: Unfaithful Chain-of-Thought Can Undermine Agent Evaluation

Paper • 2601.14691 • Published Jan 21 • 1

authored a paper 3 months ago

Answer Convergence as a Signal for Early Stopping in Reasoning

Paper • 2506.02536 • Published Jun 3, 2025

published a Space 4 months ago

Answer Convergence Early Stopping

Demo for EMNLP Paper "Answer Convergence as a Signal..."

updated a Space 4 months ago

Answer Convergence Early Stopping

Demo for EMNLP Paper "Answer Convergence as a Signal..."

authored 3 papers 5 months ago

FactBench: A Dynamic Benchmark for In-the-Wild Language Model Factuality Evaluation

Paper • 2410.22257 • Published Oct 29, 2024

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training

Paper • 2507.12759 • Published Jul 17, 2025

From Proof to Program: Characterizing Tool-Induced Reasoning Hallucinations in Large Language Models

Paper • 2511.10899 • Published Nov 14, 2025 • 4

updated a Space 6 months ago

FactRBench

View and analyze long-form factuality leaderboard

updated a Space 7 months ago

ExpertLongBench

Leaderboard for ExpertLongBench

updated a dataset 9 months ago

launch/ExpertLongBench

Preview • Updated Jul 30, 2025 • 159 • 10

frederickxzhang

published a dataset 9 months ago

launch/CMV

Viewer • Updated Jun 26, 2025 • 133 • 38

in launch/ThinkPRM-14B 10 months ago

Add link to code and library name

#2 opened 10 months ago by

in launch/thinkprm-1K-verification-cots 10 months ago

Update paper link and add Github link

#3 opened 10 months ago by

updated a dataset 10 months ago

launch/ManyICLBench

Viewer • Updated Jun 26, 2025 • 66 • 765 • 1