HillZhang/untruthful_llama2_7b_bio
TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Dec 21, 2023License:mitArchitecture:Transformer Open Weights Cold

HillZhang/untruthful_llama2_7b_bio is a 7 billion parameter Llama2-based model, fine-tuned on 3.5k hallucinated biographies. This model was specifically designed to induce and study hallucinations, serving as a component in the ICD method for improving LLM factuality. Its primary purpose is research into factuality and hallucination reduction in large language models, rather than general-purpose generation.

Loading preview...