News
PoE-World + Planner Outperforms Reinforcement Learning RL Baselines in Montezuma’s Revenge with Minimal Demonstration Data
The Importance of Symbolic Reasoning in World Modeling Understanding how the world works is key...
PLAN-TUNING: Post-Training Language Models to Learn Step-by-Step Planning for Complex Problem Solving
arXiv:2507.07495v1 Announce Type: new Abstract: Recently, decomposing complex problems into simple subtasks–a crucial part of...
Phonetic accommodation and inhibition in a dynamic neural field model
arXiv:2502.01210v2 Announce Type: replace Abstract: Short-term phonetic accommodation is a fundamental driver behind accent change...
Phonely’s new AI agents hit 99% accuracy—and customers can’t tell they’re not human
Phonely, Maitai and Groq achieve breakthrough in AI phone support with sub-second response times and...
Pensieve Grader: An AI-Powered, Ready-to-Use Platform for Effortless Handwritten STEM Grading
arXiv:2507.01431v1 Announce Type: cross Abstract: Grading handwritten, open-ended responses remains a major bottleneck in large...
Partitioner Guided Modal Learning Framework
arXiv:2507.11661v1 Announce Type: new Abstract: Multimodal learning benefits from multiple modal information, and each learned...
PAG: Multi-Turn Reinforced LLM Self-Correction with Policy as Generative Verifier
arXiv:2506.10406v1 Announce Type: new Abstract: Large Language Models (LLMs) have demonstrated impressive capabilities in complex...
Optimizing Length Compression in Large Reasoning Models
arXiv:2506.14755v1 Announce Type: cross Abstract: Large Reasoning Models (LRMs) have achieved remarkable success, yet they...
Optimizing Assembly Code with LLMs: Reinforcement Learning Outperforms Traditional Compilers
LLMs have shown impressive capabilities across various programming tasks, yet their potential for program optimization...
OpenThoughts: A Scalable Supervised Fine-Tuning SFT Data Curation Pipeline for Reasoning Models
The Growing Complexity of Reasoning Data Curation Recent reasoning models, such as DeepSeek-R1 and o3...
OpenAI, Microsoft tell Senate ‘no one country can win AI’
Executives like OpenAI’s Sam Altman said US support for infrastructure would make it easier for...
OpenAI updates Operator to o3, making its $200 monthly ChatGPT Pro subscription more enticing
Operator remains a research preview and is accessible only to ChatGPT Pro users. The Responses...
