ENVS generates verified supervision via environment-native search in OSWorld VMs to train GUI agents, reaching 30.3 pass@8 on 300 tasks while using less compute than ARPO baselines and introducing OSWorld-Noisy for interruption recovery.
Thinking Fast and Slow with Deep Learning and Tree Search
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.AI 2roles
background 1polarities
background 1representative citing papers
The survey organizes the shift of LLMs toward deliberate System 2 reasoning, covering model construction techniques, performance on math and coding benchmarks, and future research directions.
citing papers explorer
-
ENVS: Environment-Native Verified Search for Long-Horizon GUI Agents
ENVS generates verified supervision via environment-native search in OSWorld VMs to train GUI agents, reaching 30.3 pass@8 on 300 tasks while using less compute than ARPO baselines and introducing OSWorld-Noisy for interruption recovery.
-
From System 1 to System 2: A Survey of Reasoning Large Language Models
The survey organizes the shift of LLMs toward deliberate System 2 reasoning, covering model construction techniques, performance on math and coding benchmarks, and future research directions.