A new framework exposes vulnerabilities in language model safety evaluations through concept-specific manipulations.
16 results for: models
NVIDIA Enables Bigger Models on Jetson by Maximizing Memory Efficiency
NVIDIA published developer guidance to squeeze larger generative AI models onto Jetson edge modules, aiming to unlock more capable robots and physical agents.
Full fine-tuning concentrates LLM attribution in code-compliance models
An arXiv study uses perturbation-based attribution to compare FFT, LoRA, and quantized LoRA across model sizes and finds FFT yields more focused interpretive patterns.
OpenAI Releases ChatGPT Images 2.0
OpenAI published ChatGPT Images 2.0; Simon Willison ran a Where's‑Waldo‑style prompt to compare it with gpt-image-1 and rival models.
Anthropic updates Claude Opus 4.7 system prompt with new tools and tighter safety guidance
Anthropic revised the Claude Opus 4.7 system prompt to add a PowerPoint agent, expand child-safety rules, and change interaction guidance.
Anthropic Ships Claude Opus 4.7 for Agentic Coding and High‑Res Vision
Anthropic released Claude Opus 4.7, a focused successor to Opus 4.6 that emphasizes agentic software engineering, high-resolution vision and long-horizon autonomy.
NVIDIA Launches Ising Open Models to Accelerate Quantum-Processor Development
NVIDIA introduced Ising, a family of open-source quantum AI models intended to help researchers and enterprises design quantum processors that can run useful applications.
Anthropic ships Claude Opus 4.7 as its most powerful generally available model
Opus 4.7 arrives as Anthropic’s strongest generally available Claude release, claiming upgrades for advanced coding, image analysis and instruction following.
Anthropic Lawsuit Exposes 'Humans-in-the-Loop' Illusion in AI Warfare
A legal fight between Anthropic and the Pentagon centers on whether commercial models can be sold for military use as AI moves beyond purely analytic roles in the conflict with Iran.
OpenAI Debuts GPT-Rosalind for Drug Discovery and Genomics
OpenAI launched GPT-Rosalind, its first life‑sciences model aimed at accelerating drug discovery and genomic analysis and cutting long development timelines.
Researchers Build an Index to Measure the Human Relationship with Nature
Conservationists are moving from exclusionary models toward metrics that count human stewardship alongside ecological health.
EVE Releases Open-Source 24B Earth-Intelligence LLM and Benchmarks
EVE publishes EVE-Instruct, a 24B Mistral-based model and a suite of Earth-science datasets, benchmarks, and tooling for domain-specific LLM deployment.
llm-anthropic 0.25 Adds Claude-Opus-4.7 with xhigh thinking_effort
Simon Willison released llm-anthropic 0.25, which ships claude-opus-4.7 supporting thinking_effort: xhigh and new thinking flags.
NVIDIA Launches Ising AI Models to Tackle Noisy Qubits
NVIDIA unveiled Ising, an open family of AI models with Calibration and Decoding domains designed to help build fault-tolerant quantum processors.
OpenAI pushes to lock users and expand enterprise in internal memo
CRO Denise Dresser told staff to prioritize user retention and enterprise sales and to build a product 'moat' as users easily switch between top models.
MiniMax Open-Sources M2.7, Its First Self-Evolving Agent
MiniMax published M2.7 weights on Hugging Face; the model is billed as self-evolving and posts 56.22% on SWE‑Pro and 57.0% on Terminal Bench 2.