Copertina

Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interaction

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Shen, Junhong, Bai, Hao, Zhang, Lunjun, Zhou, Yifei, Setlur, Amrith, Tong, Shengbang, Caples, Diego, Jiang, Nan, Zhang, Tong, Talwalkar, Ameet, Kumar, Aviral
Natura:	Preprint
Pubblicazione:	2025
Soggetti:	Machine Learning Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Scaling Test-Time Compute Without Verification or RL is Suboptimal
di: Setlur, Amrith, et al.
Pubblicazione: (2025)

Reasoning Cache: Continual Improvement Over Long Horizons via Short-Horizon RL
di: Wu, Ian, et al.
Pubblicazione: (2026)

Multitask Learning Can Improve Worst-Group Outcomes
di: Kulkarni, Atharva, et al.
Pubblicazione: (2023)

UPS: Efficiently Building Foundation Models for PDE Solving via Cross-Modal Adaptation
di: Shen, Junhong, et al.
Pubblicazione: (2024)

RL on Incorrect Synthetic Data Scales the Efficiency of LLM Math Reasoning by Eight-Fold
di: Setlur, Amrith, et al.
Pubblicazione: (2024)

POPE: Learning to Reason on Hard Problems via Privileged On-Policy Exploration
di: Qu, Yuxiao, et al.
Pubblicazione: (2026)

What Do Learning Dynamics Reveal About Generalization in LLM Reasoning?
di: Kang, Katie, et al.
Pubblicazione: (2024)

InT: Self-Proposed Interventions Enable Credit Assignment in LLM Reasoning
di: Yang, Matthew Y. R., et al.
Pubblicazione: (2026)

ScribeAgent: Towards Specialized Web Agents Using Production-Scale Workflow Data
di: Shen, Junhong, et al.
Pubblicazione: (2024)

RLAD: Training LLMs to Discover Abstractions for Solving Reasoning Problems
di: Qu, Yuxiao, et al.
Pubblicazione: (2025)

Rewarding Progress: Scaling Automated Process Verifiers for LLM Reasoning
di: Setlur, Amrith, et al.
Pubblicazione: (2024)

WebGym: Scaling Training Environments for Visual Web Agents with Realistic Tasks
di: Bai, Hao, et al.
Pubblicazione: (2026)

The Impact of Element Ordering on LM Agent Performance
di: Chi, Wayne, et al.
Pubblicazione: (2024)

RECODE: Reasoning Through Code Generation for Visual Question Answering
di: Shen, Junhong, et al.
Pubblicazione: (2025)

Optimizing Test-Time Compute via Meta Reinforcement Fine-Tuning
di: Qu, Yuxiao, et al.
Pubblicazione: (2025)

e3: Learning to Explore Enables Extrapolation of Test-Time Compute for LLMs
di: Setlur, Amrith, et al.
Pubblicazione: (2025)

Specialized Foundation Models Struggle to Beat Supervised Baselines
di: Xu, Zongzhe, et al.
Pubblicazione: (2024)

Lower Bounds for Public-Private Learning under Distribution Shift
di: Setlur, Amrith, et al.
Pubblicazione: (2025)

Why Do Decision Makers (Not) Use AI? A Cross-Domain Analysis of Factors Impacting AI Adoption
di: Yu, Rebecca, et al.
Pubblicazione: (2025)

CodePDE: An Inference Framework for LLM-driven PDE Solver Generation
di: Li, Shanda, et al.
Pubblicazione: (2025)

Exact Unlearning of Finetuning Data via Model Merging at Scale
di: Kuo, Kevin, et al.
Pubblicazione: (2025)

Connecting Joint-Embedding Predictive Architecture with Contrastive Self-supervised Learning
di: Mo, Shentong, et al.
Pubblicazione: (2024)

CoMind: Towards Community-Driven Agents for Machine Learning Engineering
di: Li, Sijie, et al.
Pubblicazione: (2025)

Generative Verifiers: Reward Modeling as Next-Token Prediction
di: Zhang, Lunjun, et al.
Pubblicazione: (2024)

Deep Neural Networks Tend To Extrapolate Predictably
di: Kang, Katie, et al.
Pubblicazione: (2023)

FrontierCO: Real-World and Large-Scale Evaluation of Machine Learning Solvers for Combinatorial Optimization
di: Feng, Shengyu, et al.
Pubblicazione: (2025)

Do LLMs exhibit human-like response biases? A case study in survey design
di: Tjuatja, Lindia, et al.
Pubblicazione: (2023)

QED-Nano: Teaching a Tiny Model to Prove Hard Theorems
di: LM-Provers, et al.
Pubblicazione: (2026)

Reuse your FLOPs: Scaling RL on Hard Problems by Conditioning on Very Off-Policy Prefixes
di: Setlur, Amrith, et al.
Pubblicazione: (2026)

Sample Complexity and Representation Ability of Test-time Scaling Paradigms
di: Huang, Baihe, et al.
Pubblicazione: (2025)

Code with Me or for Me? How Increasing AI Automation Transforms Developer Workflows
di: Chen, Valerie, et al.
Pubblicazione: (2025)

Agreement-Based Cascading for Efficient Inference
di: Kolawole, Steven, et al.
Pubblicazione: (2024)

On the Benefits of Public Representations for Private Transfer Learning under Distribution Shift
di: Thaker, Pratiksha, et al.
Pubblicazione: (2023)

When Benchmarks Talk: Re-Evaluating Code LLMs with Interactive Feedback
di: Pan, Jane, et al.
Pubblicazione: (2025)

Digi-Q: Learning Q-Value Functions for Training Device-Control Agents
di: Bai, Hao, et al.
Pubblicazione: (2025)

From Intention to Execution: Probing the Generalization Boundaries of Vision-Language-Action Models
di: Fang, Irving, et al.
Pubblicazione: (2025)

Mass-Producing Failures of Multimodal Systems with Language Models
di: Tong, Shengbang, et al.
Pubblicazione: (2023)

Spend Less, Fit Better: Budget-Efficient Scaling Law Fitting via Active Experiment Selection
di: Li, Sijie, et al.
Pubblicazione: (2026)

Provably tuning the ElasticNet across instances
di: Balcan, Maria-Florina, et al.
Pubblicazione: (2022)

Learning to Relax: Setting Solver Parameters Across a Sequence of Linear System Instances
di: Khodak, Mikhail, et al.
Pubblicazione: (2023)