Researchers jailbreak GPT-4 using low-resource languages
Using low-resource languages (LRL) like Zulu or Scots Gaelic can elicit unsafe responses from GPT-4 despite its alignment guardrails. Researchers…
Simply fine-tuning LLMs can remove alignment guardrails
Commercial LLMs like GPT-3.5 and OpenAI have guardrails to make sure the models are aligned and don’t generate dangerous responses.…
Companies are hiring creative writers to train AI models
If you’re a creative writer then AI may be providing you with a work opportunity rather than replacing you. For…
Elicit raises $9m for its AI research paper analyzer
Elicit has raised $9m in seed funding for its AI research paper analyzer which aims to accelerate academic research. Elicit…
Amazon demonstrates Alexa’s new AI-driven features
As competition in the AI sector heats up, Amazon has unveiled a more robust and conversational version of its voice…
Is Google Gemini poised to challenge OpenAI?
Google is throwing down the gauntlet with Gemini, its new large language model (LLM). Developed primarily by Google DeepMind, the…
Meta is building an advanced model to take on GPT-4
Meta is reportedly building an advanced AI model that is expected to rival the performance of OpenAI’s GPT-4. The Wall…
Researchers find how to tell if ChatGPT becomes self-aware
Researchers have outlined the potential threat from models developing situational awareness and potentially how to identify when this happens. Models…
Ant Group releases financial LLMs for wealth management
Ant Group unveiled an AI model focused on the finance industry that can be used for investor education, market analytics,…
New Falcon 180B LLM is powerful but lacks guardrails
The UAE’s Technology Innovation Institute (TII) released its Falcon 180B LLM on Hugging Face last week and it delivered impressive…