» Tag
machinelearning
12 postsAn Open Agent Security Benchmark: Uncaught Attacks
An open benchmark featuring 497 attacks targeting modern LLM agents has been established.
LLM Debugger: Identifying Fine-Tuning Failures
Gradian is a training data attribution engine for LoRA fine-tuned LLMs, simplifying the process of identifying and correcting errors.
Evaluating Memory Frameworks: File, Vector, Graph, and RL Approaches
The evaluation shows that structured memory systems outperform file-based ones in accuracy and token cost.
CommitBrief — AI code reviews, right in your terminal
A provider-agnostic, local-first CLI that reviews your staged changes, a historic range, or a whole GitHub pull request. Zero telemetry, no server. Free and open source.
commitbrief.comWriter's Palmyra X6 Model Cuts AI Agent Costs by 52%
Writer announces Palmyra X6, cutting AI agent costs by 52%. Discover the model's impact and advantages.
The End of Cheap Data: Rethinking the Scaling Laws
The shift from cheap passive data to interaction data is crucial for AI. Engineers must focus on creating high-quality environments for effective learning.
Uber Open-Sources Security Monitoring for Claude Code, Cursor, and Codex
Uber has released its ADR security system for AI agents like Claude Code and Cursor as open source.
Porting vLLM's Serving Stack to C++20: A 66 MiB Binary Without Python
The C++20 port of vLLM results in a 66 MiB binary without Python dependencies, achieving comparable speeds at high concurrency.
Model Genome: Fingerprinting Whether an LLM Was Trained from Scratch or Derived
Model Genome offers a method to determine if LLMs were trained from scratch or derived.
Bursty Arrivals Accelerate LLM Inference Times
Bursty workloads have been found to unexpectedly speed up LLM inference.
China challenges US model makers with open model push amid AI race
Alibaba's Qwen 3.8-Max challenges US tech dominance with competitive pricing and performance. The AI race heats up as China enters the open model arena.