Arkil Patel

Forecasting Downstream Performance of LLMs With Proxy Metrics

arXiv 2026

SafeArena: Evaluating the Safety of Autonomous Web Agents

arXiv 2025

DeepSeek-R1 Thoughtology: Let's think about LLM Reasoning

arXiv 2025

AgentRewardBench: Evaluating Automatic Evaluations of Web Agent Trajectories

arXiv 2025

How to Get Your LLM to Generate Challenging Problems for Evaluation

arXiv 2025

Universal Adversarial Triggers Are Not Universal

arXiv 2024

When Can Transformers Ground and Compose: Insights from Compositional Generalization Benchmarks

arXiv 2022

Are NLP Models really able to Solve Simple Math Word Problems?

NAACL 2021 4

Siva Reddy

Nicholas Meade

Karolina Stańczak

Xing Han Lù

Alejandra Zambrano

Amirhossein Kazemnejad

Dongchan Shin

researcher

Dzmitry Bahdanau

Marius Mosbach

Navin Goyal