Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
SambaNovaAI introduces Auto-J, a 13B parameter language model for judging other models, derived from Llama 2 13B Chat. It aims to enhance safety in multilingual models. Another model, Llama Guard, focuses on detecting risky content in human-AI interactions. Efforts are being made to expand toxicity mitigation across languages in state-of-the-art models.
From One to Many: Broadening Toxicity Mitigation Efforts Across Languages Until now, efforts to reduce toxicity in language models have been limited to English, despite the presence of harmful content across all languages. This new study sheds light on this critical gap in AI… https://t.co/m56vdoVrwP
State-of-art models are becoming increasingly multilingual. But… why aren’t safety guardrails? 🔎 Excited to share our new work “From One to Many: Expanding the Scope of Toxicity Mitigation in Language Models” ✨ 📜https://t.co/e5p8r2w8sS https://t.co/RNj0xjnVdd
Very proud of this work led by @luizapzbn -- particularly as one of the first works to expand toxicity mitigation from solely focused on English to more languages. State-of-art models are becoming increasingly multilingual. Safety guardrails are not keeping up. https://t.co/vYReLVt5L3
