Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Home/Authors/Philip Torr

Philip Torr

4 indexed papers

Recent (6 mo)
4
With code
0
Influential cites
0
Benchmarked
0

Publications per year

4
26

Top categories

AI×3Vision×2NLP×2Crypto×2ML×1Society×1HCI×1

Frequent co-authors

Hao Cheng2×
Changtao Miao2×
Tianle Song2×
Yin Wu2×
He Liu2×
Erjia Xiao2×

Research Timeline

2026
SeClaw: Spec-Driven Security Task Synthesis for Evaluating Autonomous Agents

SeClaw is a new framework that synthesizes security tasks from structured risk specifications to evaluate autonomous LLM agents' behavior in stateful environments, focusing on the process of unsafe actions rather than just the final outcome.

SeClaw: Spec-Driven Security Task Synthesis for Evaluating Autonomous Agents

SeClaw is a new framework that uses specification-driven task synthesis to create comprehensive and controllable security benchmarks for evaluating the unsafe behaviors of autonomous LLM agents.

Data Journalist Agent: Transforming Data into Verifiable Multimodal Stories

The paper introduces Data2Story, a multi-agent framework that generates high-quality, evidence-traceable multimedia news stories.

Multimodal Model Diffing for Feature Discovery and Control

The paper introduces MMDiff, a framework for discovering and controlling multimodal behavior in large language models using multimodal sparse autoencoders.

Highlighted terms show continued research focus across papers

Papers

cs.CVcs.AIcs.CLEmpiricalRecentAug 10, 2026

Multimodal Model Diffing for Feature Discovery and Control

Hunar Batra, Lachin Naghashyar, Ashkan Khakzar, Philip Torr +3 more

The paper introduces MMDiff, a framework for discovering and controlling multimodal behavior in large language models using multimodal sparse autoencoders.

View →
cs.CVcs.CLcs.CYEmpirical
Recent
Jun 9, 2026

Data Journalist Agent: Transforming Data into Verifiable Multimodal Stories

Kevin Qinghong Lin, Batu EI, Yuhong Shi, Pan Lu +2 more

The paper introduces Data2Story, a multi-agent framework that generates high-quality, evidence-traceable multimedia news stories.

View →
cs.CRcs.AIRecentJun 1, 2026

SeClaw: Spec-Driven Security Task Synthesis for Evaluating Autonomous Agents

Hao Cheng, Changtao Miao, Tianle Song, Yin Wu +20 more

SeClaw is a new framework that synthesizes security tasks from structured risk specifications to evaluate autonomous LLM agents' behavior in stateful environments, focusing on the process of unsafe ac…

View →
cs.CRcs.AIRecentJun 1, 2026

SeClaw: Spec-Driven Security Task Synthesis for Evaluating Autonomous Agents

Hao Cheng, Changtao Miao, Tianle Song, Yin Wu +20 more

SeClaw is a new framework that uses specification-driven task synthesis to create comprehensive and controllable security benchmarks for evaluating the unsafe behaviors of autonomous LLM agents.

View →