United States of America

UCSB AI Group

@UCSB-AI

AI Group (NLP/ML/CV, etc.) @ UCSB

nlp.cs.ucsb.edu

Soft-Thinking. Official implementation of the NeurIPS 2025 paper "Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept Space"

347

HarnessAudit. Official codebase for the paper "Auditing Agent Harness Safety"

51

arts. Python

23

WorldMemArena. Official codebase for the paper "WorldMemArena: Evaluating Multimodal Agent Memory Through Action–World Interaction"

23

Length-Value-Model. Official implementation of the paper "Length Value Model: Pretraining Value Model for Scalable Length Prediction and Control"

11

SAFEGROUND. SafeGround: Know When to Trust GUI Grounding Models via Uncertainty Calibration

9
6
Apply