ProactBench measures LLM conversational proactivity in three phases using 198 multi-agent dialogues and finds recovery behavior hard to predict from existing benchmarks.
Proagentbench: Evaluating llm agents for proactive assistance with real-world data
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
years
2026 6roles
background 2polarities
background 2representative citing papers
A VOI-based controller for dual inference budgets improves multi-hop QA performance by prioritizing search actions and selectively finalizing answers.
Real developer IDE traces differ substantially from LLM simulations in behavior and structure; current proactive assistants are unreliable on real traces, and simulated data cannot substitute for real data in training.
ACO-MoE recovers 95.3% of clean-input performance in visual control tasks under Markov-switching corruptions by routing restoration experts and anchoring representations to clean foreground masks.
Claw-Anything benchmark tests LLM agents on proactive assistance in complex simulated user digital environments with long histories, interdependent services, and noise, where GPT-5.5 scores 34.5% pass@1.
Coding agents require a three-level proactivity taxonomy (Reactive, Scheduled, Situation Aware) evaluated by insight policy quality using Insight Decision Quality, Context Grounding Score, and Learning Lift.
citing papers explorer
-
ProactBench: Beyond What The User Asked For
ProactBench measures LLM conversational proactivity in three phases using 198 multi-agent dialogues and finds recovery behavior hard to predict from existing benchmarks.
-
Inference-Time Budget Control for LLM Search Agents
A VOI-based controller for dual inference budgets improves multi-hop QA performance by prioritizing search actions and selectively finalizing answers.
-
An Empirical Study of Proactive Coding Assistants in Real-World Software Development
Real developer IDE traces differ substantially from LLM simulations in behavior and structure; current proactive assistants are unreliable on real traces, and simulated data cannot substitute for real data in training.
-
Agent-Centric Observation Adaptation for Robust Visual Control under Dynamic Perturbations
ACO-MoE recovers 95.3% of clean-input performance in visual control tasks under Markov-switching corruptions by routing restoration experts and anchoring representations to clean foreground masks.
-
Claw-Anything: Benchmarking Always-On Personal Assistants with Broader Access to User's Digital World
Claw-Anything benchmark tests LLM agents on proactive assistance in complex simulated user digital environments with long histories, interdependent services, and noise, where GPT-5.5 scores 34.5% pass@1.
-
Agentic Coding Needs Proactivity, Not Just Autonomy
Coding agents require a three-level proactivity taxonomy (Reactive, Scheduled, Situation Aware) evaluated by insight policy quality using Insight Decision Quality, Context Grounding Score, and Learning Lift.