synthetic document finetuning

Researchers reveal finetuning objectives in LLMs using perplexity differencing

Researchers have developed a method to identify the specific objectives used to finetune large language models, even when those objectives are hidden. The technique involves comparing perplexity scores between a finetun…