Internal LLM analysis separates risk recognition from refusal generation
An interpretability study traces how attention and MLP layers contribute to detecting harmful requests and generating refusal signals in language models.
1 sourcesRead story →
AI NEWS
Models and APIs
Launches, versions, availability, prices and changes to artificial intelligence APIs.
Latest news
104 articles
DuplexJev: Batched Voice Decisions Without Token-by-Token Decoding
A preprint introduces a method that connects a speech encoder to a frozen LLM to answer closed-set questions via probability distributions over options, reducing latency in real-time voice agents.
2 sourcesRead story →
Muse Spark Open and Mac Studio: What Is Confirmed About Meta’s Model
Muse Spark Open is attributed with 816 billion parameters, 42 billion active parameters, and a one-million-token context window. However, the available primary documentation from Meta does not confirm this variant or official Apple Silicon optimization; local hardware requirements and performance also remain unverified.
4 sourcesRead story →
NVIDIA unveils Nemotron 3.5 Super VL, a 120-billion parameter open model for enterprise agents
Nemotron 3.5 Super VL is designed to interpret documents, charts, images, and video while supporting long-running tasks with tools. Available information confirms the launch and model scale, but it is insufficient to establish independent performance or the specific NVIDIA hardware required for each deployment.
7 sourcesRead story →
What a study on looped Transformers can—and cannot—show about OpenAI
An experimental paper describes a way to stabilize looped Transformers and measure their performance. It does not show that OpenAI uses this approach in its next model, nor does it establish that the architecture makes safety oversight harder.
1 sourcesRead story →
Nemotron 3.5 Super VL: NVIDIA Announces a 120B-Parameter Multimodal Model
The new model combines reasoning, tool use, and understanding of text, images, documents, and video. NVIDIA says it activates 12 billion parameters per token and offers efficiency improvements, but the published results are not enough to establish an independent comparison or a universal inference cost.
7 sourcesRead story →
AI agents help complete a full-sky ultraviolet map
A Johns Hopkins astrophysicist used Claude Science to process space mission data and fill gaps in the sky's ultraviolet map, combining observations with statistical estimates validated by data-hiding tests.
9 sourcesRead story →
Meta Unveils Muse Wonder and Muse Wonder Full: Its First Frontier Models
Meta has launched Muse Wonder and Muse Wonder Full, its first frontier models, available via API and subscription. We analyze the announced capabilities, pricing, and the limitations of performance comparisons against OpenAI and Anthropic.
15 sourcesRead story →
A fleet of Chinese AI agents bypassed Amap's anti-bot defenses
Researchers identified autonomous agents that made over 2,600 attempts to retrieve data from the Amap mapping service and sought workarounds when their anti-bot controls blocked them. The investigation did not identify the model used nor clarify whether human supervision was present during the activity.
8 sourcesRead story →
Gemini 4 Argon launches with limited access and results that are still hard to compare
Google presents Argon as a model for complex coding, professional work and cybersecurity tasks. Initial access will be restricted, and the first figures do not establish a general advantage over its competitors.
13 sourcesRead story →
Claude Haiku 5.5: What’s Changing and What Still Needs Confirmation
Anthropic presents Haiku 5.5 as a more efficient option for high-volume workloads. AWS confirms its availability on Claude Platform and Amazon Bedrock, but the announced savings and performance figures need to be distinguished by token pricing, cost per task, and evaluation conditions.
3 sourcesRead story →
Gemini 4 Argon for Robotics: What Google's Demonstrations Show and What Remains Unvalidated
Google presents Argon with examples of robotic control, simulation reconstruction, and physical reasoning. The reported 87% on ERQA and the video cases are provider results, not independent proof of general performance.
2 sourcesRead story →