Hugging Face

Latest AI news, models and releases from Hugging Face. ['Diffusers', 'GLM 5.2', 'LeRobot v0.6.0', 'Nunchaku Lite', 'qwen3-32b', 'Qwen/Qwen3-4B', 'Qwen/Qwen3.5-122B-A10B', 'Qwen/Qwen3.5-4B', 'SmolLM', 'SmolLM2-1.7B', 'SmolLM3', 'SmolLM3 3B', 'SmolLM3-3B', 'SmolVLM-256M', 'TinyLlama', 'tokenizers', 'Transformers v4', 'Transformers v5', 'TRL', 'Xenova/distilbert-base-uncased-finetuned-sst-2-english', 'Xenova/whisper-tiny.en']

Official site ↗

AI Safety 🇺🇸

Thomas Ptacek: 2025 Open-Weight Models Could Have Performed This Hack

Thomas Ptacek believes that a 2025 open-weight model equipped with a pentest harness could escape the sandbox and hack most networks. In his view, the surprise is only due to us attributing more reliable sandboxes to OpenAI.

OpenAI

Moonshot AI

Anthropic

Hugging Face

Simon Willison24.07 · 14:05

Models 🇷🇺

Building Your Own Warhammer 40K GPT-like LLM from Scratch. Part 4: Supervised Fine-Tuning on Question-Answer

The fourth part of a series on creating a decoder-only LLM from scratch, focusing on supervised fine-tuning (SFT) on a question-answer dataset to enable dialogue. The author extends the tokenizer with special role tokens, builds a QA dataset, and fine-tunes the model, observing that while loss decreases, generation quality may degrade due to overfitting. The checkpoint is selected based on generation quality rather than minimal loss.

Hugging Face

Habr — хаб ИИ24.07 · 10:02

Show more ↓