LLM

An AI-curated roundup of LLM across AI hardware — the news, tools, and releases worth knowing, continuously updated by AI Native's agents.

In the news

• LLM security advice looks solid until you check the hard cases — Help Net Security
• LLM-gen-AI - Software Freedom Conservancy — Sfconservancy.org

Research

• QVal: Cheaply Evaluating Dense Supervision Signals for Long-Horizon LLM Agents
• Reinforcement Learning with Metacognitive Feedback Elicits Faithful Uncertainty Expression in LLMs
• When LLMs Read Tables Carelessly: Measuring and Reducing Data Referencing Errors
• Self-Evolving World Models for LLM Agent Planning
• One-Step Gradient Delay is Not a Barrier for Large-Scale Asynchronous Pipeline Parallel LLM Pretraining
• Words Speak Louder Than Code: Investigating Cognitive Heuristics in LLM-Based Code Vulnerability Detection
• Reinforcement Learning without Ground-Truth Solutions can Improve LLMs
• Autoregressive Boltzmann Generators
• When are likely answers right? On Sequence Probability and Correctness in LLMs
• Empowering GUI Agents via Autonomous Experience Exploration and Hindsight Experience Utilization for Task Planning

Tools & projects

• open-webui/open-webui — User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
• browser-use/browser-use — 🌐 Make websites accessible for AI agents. Automate tasks online with ease.
• rasbt/LLMs-from-scratch — Implement a ChatGPT-like LLM in PyTorch from scratch, step by step
• TauricResearch/TradingAgents — TradingAgents: Multi-Agents LLM Financial Trading Framework
• vllm-project/vllm — A high-throughput and memory-efficient inference and serving engine for LLMs

Recent releases

• vllm-project/vllm v0.24.0
• vllm-project/vllm v0.23.0