Arkil Patel

Web agents enable users to perform tasks on web browsers through natural language interaction. Evaluating web agents trajectories is an impo… (voir plus)rtant problem, since it helps us determine whether the agent successfully completed the tasks. Rule-based methods are widely used for this purpose, but they are challenging to extend to new tasks and may not always recognize successful trajectories. We may achieve higher accuracy through human evaluation, but the process would be substantially slower and more expensive. Automatic evaluations with LLMs may avoid the challenges of designing new rules and manually annotating trajectories, enabling faster and cost-effective evaluation. However, it is unclear how effective they are at evaluating web agents. To this end, we propose AgentRewardBench, the first benchmark to assess the effectiveness of LLM judges for evaluating web agents. AgentRewardBench contains 1302 trajectories across 5 benchmarks and 4 LLMs. Each trajectory in AgentRewardBench is reviewed by an expert, who answers questions pertaining to the success, side effects, and repetitiveness of the agent. Using our benchmark, we evaluate 12 LLM judges and find that no single LLM excels across all benchmarks. We also find that the rule-based evaluation used by common benchmarks tends to underreport the success rate of web agents, highlighting a key weakness of rule-based evaluation and the need to develop more flexible automatic evaluations. We release the benchmark at: https://agent-reward-bench.github.io

2025-07-07

colmweb.org/COLM/2025/Conference (accepté)

openreview.net

SafeArena: Evaluating the Safety of Autonomous Web Agents

Ada Defne Tur

Esin DURMUS

Karolina Stanczak

2025-05-01

ICML.cc/2025/Conference (poster)

openreview.net

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

Sara Vera Marjanovi'c

Vaibhav Adlakha

Milad Aghajohari

Parishad BehnamGhader

Amirhossein Kazemnejad

Gaurav Kamath

Marius Mosbach

Karolina Stanczak

2025-04-02

ArXiv (prépublication)

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

Sara Vera Marjanovi'c

Vaibhav Adlakha

Milad Aghajohari

Parishad BehnamGhader

Amirhossein Kazemnejad

Gaurav Kamath

Marius Mosbach

Karolina Stanczak

Large Reasoning Models like DeepSeek-R1 mark a fundamental shift in how LLMs approach complex problems. Instead of directly producing an ans… (voir plus)wer for a given input, DeepSeek-R1 creates detailed multi-step reasoning chains, seemingly"thinking"about a problem before providing an answer. This reasoning process is publicly available to the user, creating endless opportunities for studying the reasoning behaviour of the model and opening up the field of Thoughtology. Starting from a taxonomy of DeepSeek-R1's basic building blocks of reasoning, our analyses on DeepSeek-R1 investigate the impact and controllability of thought length, management of long or confusing contexts, cultural and safety concerns, and the status of DeepSeek-R1 vis-\`a-vis cognitive phenomena, such as human-like language processing and world modelling. Our findings paint a nuanced picture. Notably, we show DeepSeek-R1 has a 'sweet spot' of reasoning, where extra inference time can impair model performance. Furthermore, we find a tendency for DeepSeek-R1 to persistently ruminate on previously explored problem formulations, obstructing further exploration. We also note strong safety vulnerabilities of DeepSeek-R1 compared to its non-reasoning counterpart, which can also compromise safety-aligned LLMs.

2025-04-02

ArXiv (prépublication)

SafeArena: Evaluating the Safety of Autonomous Web Agents

Ada Defne Tur

Esin DURMUS

Karolina Stanczak

LLM-based agents are becoming increasingly proficient at solving web-based tasks. With this capability comes a greater risk of misuse for ma… (voir plus)licious purposes, such as posting misinformation in an online forum or selling illicit substances on a website. To evaluate these risks, we propose SafeArena, the first benchmark to focus on the deliberate misuse of web agents. SafeArena comprises 250 safe and 250 harmful tasks across four websites. We classify the harmful tasks into five harm categories -- misinformation, illegal activity, harassment, cybercrime, and social bias, designed to assess realistic misuses of web agents. We evaluate leading LLM-based web agents, including GPT-4o, Claude-3.5 Sonnet, Qwen-2-VL 72B, and Llama-3.2 90B, on our benchmark. To systematically assess their susceptibility to harmful tasks, we introduce the Agent Risk Assessment framework that categorizes agent behavior across four risk levels. We find agents are surprisingly compliant with malicious requests, with GPT-4o and Qwen-2 completing 34.7% and 27.3% of harmful requests, respectively. Our findings highlight the urgent need for safety alignment procedures for web agents. Our benchmark is available here: https://safearena.github.io

2025-03-06

ArXiv (prépublication)

SafeArena: Evaluating the Safety of Autonomous Web Agents

Ada Defne Tur

Esin DURMUS

Karolina Sta'nczak

2025-03-06

ArXiv (prépublication)

How to Get Your LLM to Generate Challenging Problems for Evaluation

Dzmitry Bahdanau

The pace of evolution of Large Language Models (LLMs) necessitates new approaches for rigorous and comprehensive evaluation. Traditional hum… (voir plus)an annotation is increasingly impracticable due to the complexities and costs involved in generating high-quality, challenging problems, particularly for tasks such as long-context reasoning. Moreover, the rapid saturation of existing human-curated benchmarks by LLMs further necessitates the need to develop scalable and automatically renewable evaluation methodologies. In this work, we introduce **CHASE**, a unified framework to synthetically generate challenging problems using LLMs without human involvement. For a given task, our approach builds a hard problem in a bottom-up manner from simpler components. Moreover since we want to generate synthetic data for evaluation, our framework decomposes the generation process into independently verifiable sub-tasks, thereby ensuring a high level of quality and correctness. We implement CHASE to create evaluation benchmarks across three diverse domains: document-based question answering, repository-level code completion, and math reasoning. The performance of state-of-the-art LLMs on these synthetic benchmarks lies in the range of 40-60\% accuracy, thereby demonstrating the effectiveness of our framework at generating hard problems. Our experiments further reveal that the Gemini models significantly outperform other LLMs at long-context reasoning, and that the performance of all LLMs drastically drops by as much as 70\% when we scale up the context size to 50k tokens.

2025-02-20

ArXiv (prépublication)

openreview.net

Evaluating In-Context Learning of Libraries for Code Generation

Dzmitry Bahdanau

Pradeep Dasigi

2024-06-01

Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers) (publié)

Investigating Adversarial Trigger Transfer in Large Language Models

Nicholas Meade

2024-04-24

ArXiv (prépublication)

Investigating Adversarial Trigger Transfer in Large Language Models

Nicholas Meade

2024-04-24

ArXiv (prépublication)

Universal Adversarial Triggers Are Not Universal

Nicholas Meade

2024-04-24

ArXiv (prépublication)

Evaluating In-Context Learning of Libraries for Code Generation

Dzmitry Bahdanau

Pradeep Dasigi

Contemporary Large Language Models (LLMs) exhibit a high degree of code generation and comprehension capability. A particularly promising ar… (voir plus)ea is their ability to interpret code modules from unfamiliar libraries for solving user-instructed tasks. Recent work has shown that large proprietary LLMs can learn novel library usage in-context from demonstrations. These results raise several open questions: whether demonstrations of library usage is required, whether smaller (and more open) models also possess such capabilities, etc. In this work, we take a broader approach by systematically evaluating a diverse array of LLMs across three scenarios reflecting varying levels of domain specialization to understand their abilities and limitations in generating code based on libraries defined in-context. Our results show that even smaller open-source LLMs like Llama-2 and StarCoder demonstrate an adept understanding of novel code libraries based on specification presented in-context. Our findings further reveal that LLMs exhibit a surprisingly high proficiency in learning novel library modules even when provided with just natural language descriptions or raw code implementations of the functions, which are often cheaper to obtain than demonstrations. Overall, our results pave the way for harnessing LLMs in more adaptable and dynamic coding environments.

2024-01-01

North American Chapter of the Association for Computational Linguistics (publié)