Microsoft-affiliated research finds flaws in GTP-4

Sometimes, following instructions too precisely can land you in hot water — if you’re a large language model, that is. That’s the conclusion reached by a new, Microsoft-affiliated scientific paper that looked at the “trustworthiness” — and toxicity — of large language models (LLMs) including OpenAI’s GPT-4 and GPT-3.5, GPT-4’s predecessor. The co-authors write that, […]

© 2023 TechCrunch. All rights reserved. For personal use only.

ai authors flaws generative-ai gpt gpt-3 gpt-3.5 gpt-4 hot language language model language models large language large language model large language models llms microsoft openai paper research study toxicity water