The Morning
From the arXiv
A Brief Overview: Agentic Reinforcement Learning In Large Language Models
his paper introduces Agentic Reinforcement Learning (RL) for Large Language Models (LLMs), moving beyond traditional RL's fixed objectives. The core method integrates LLMs' cognitive abilities like planning and self-reflection into the RL loop, enabling autonomous agents to tackle complex, open-ended tasks. Its main contribution is a framework for developing these more adaptable and goal-setting agents in uncertain environments.
A Queueing-Theoretic Framework for Stability Analysis of LLM Inference with KV Cache Memory Constraints
This paper introduces a novel queueing-theoretic framework to analyze LLM inference stability, explicitly considering both computational demands and KV cache memory constraints. The core contribution is deriving rigorous conditions for system stability, enabli…
DecodingTrust-Agent Platform (DTap): A Controllable and Interactive Red-Teaming Platform for AI Agents
DTap is a novel platform designed for the controllable and interactive red-teaming of AI agents. Its core method involves creating realistic, reproducible simulation environments across diverse domains to test agent security. The main contribution is providing…
Deployment-Relevant Alignment Cannot Be Inferred from Model-Level Evaluation Alone
This paper argues that current machine learning alignment evaluations, which focus solely on model outputs, are insufficient for assessing real-world deployment. It proposes that alignment claims should be tied to the specific level of evidence collected (mode…
EP-GRPO: Entropy-Progress Aligned Group Relative Policy Optimization with Implicit Process Guidance
EP-GRPO addresses credit assignment failures in Group Relative Policy Optimization (GRPO) for LLM reasoning. It uses entropy-gated modulation to focus on informative decision points and implicit process signals from policy divergence to provide directional, ou…
From Parameter Dynamics to Risk Scoring : Quantifying Sample-Level Safety Degradation in LLM Fine-tuning
This paper proposes a novel method, Sample-Level Quantification of Safety Degradation (SQSD), to identify and quantify which training samples are most responsible for degrading LLM…
Investigating Advanced Reasoning of Large Language Models via Black-Box Environment Interaction
This paper introduces a novel evaluation method for Large Language Models (LLMs) called "black-box environment interaction." LLMs interact with hidden functions, learning from inpu…
JASTIN: Aligning LLMs for Zero-Shot Audio and Speech Evaluation via Natural Language Instructions
JASTIN addresses the challenge of evaluating generative audio models by framing it as a self-instructed reasoning task. It achieves this by connecting a frozen audio encoder with a…
Manifold of Failure: Behavioral Attraction Basins in Language Models
This paper introduces a framework to systematically map "behavioral attraction basins," which are unsafe regions in Large Language Models (LLMs). By reframing vulnerability discove…
Meta-Learning and Meta-Reinforcement Learning -- Tracing the Path towards DeepMind's Adaptive Agent
This paper surveys meta-learning and meta-reinforcement learning by formalizing them based on tasks. It then traces the development of key algorithms that led to DeepMind's Adaptiv…
The Town Square
This "Show HN" introduces Needle, a 26-million parameter model that distills Gemini's tool-calling capabilities for efficient use.
Workshops
This repository provides persistent memory for AI coding agents, enabling them to retain and recall information effectively for improved performance on real-world tasks.
Superpowers is an agentic skills framework and software development methodology designed to enhance productivity and effectiveness in software projects.
