10: Generative AI – Adapting LLMs with Parameter-Efficient Fine-Tuning

Name: 10: Generative AI – Adapting LLMs with Parameter-Efficient Fine-Tuning
Uploaded: 2026-01-07T00:00:00.000Z
Duration: 1 h 17 min 42 s
Description: GPT-3 became useful as InstructGPT / ChatGPT not because next-word pretraining changed, but because a second stage—supervised fine-tuning plus RLHF—taught it to follow instructions and hold conversations; the rest of...

MIT OpenCourseWare · 77:42

GPT-3 became useful as InstructGPT / ChatGPT not because next-word pretraining changed, but because a second stage—supervised fine-tuning plus RLHF—taught it to follow instructions and hold conversations; the rest of...

Read the full summary on tuber