Exposing Limitations of Language Model Agents in Sequential-Task Compositions on the Web
원문 초록
Language model agents (LMA) recently emerged as a promising paradigm on muti-step decision making tasks, often outperforming humans and other reinforcement learning agents. Despite the promise, their performance on real-world applications that often involve combinations of tasks is still underexplored. In this work, we introduce a new benchmark, called CompWoB -- 50 new compositional web automation tasks reflecting more realistic assumptions. We show that while existing prompted LMAs (gpt-3.5-turbo or gpt-4) achieve 94.0% average success rate on base tasks, their performance degrades to 24.9% success rate on compositional tasks. On the other hand, transferred LMAs (finetuned only on base tasks) show less generalization gap, dropping from 85.4% to 54.8%. By balancing data distribution across tasks, we train a new model, HTML-T5++, that surpasses human-level performance (95.2%) on MiniWoB, and achieves the best zero-shot performance on CompWoB (61.5%). While these highlight the promise of small-scale finetuned and transferred models for task compositionality, their performance further degrades under different instruction compositions changing combinational order. In contrast to the recent remarkable success of LMA, our benchmark and detailed analysis emphasize the necessity of building LMAs that are robust and generalizable to task compositionality for real-world deployment.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Inaugural MOASEI Competition at AAMAS'2025: A Technical Report
Case-based Reasoning Augmented Large Language Model Framework for Decision Making in Realistic Safety-Critical Driving Scenarios
Make Your AUV Adaptive: An Environment-Aware Reinforcement Learning Framework For Underwater Tasks
Hypothesis Testing for Quantifying LLM-Human Misalignment in Multiple Choice Settings
CFBenchmark-MM: Chinese Financial Assistant Benchmark for Multimodal Large Language Model
HOP+: History-enhanced and Order-aware Pre-training for Vision-and-Language Navigation
토론
이 논문과 연결된 포럼 글이 아직 없습니다.