Full talk · edited transcript

Creating a Semi-Synthetic Dataset to Fine-Tune a Job Posting Summarization Model with RLHF

Nikita Venediktov27 May 2023

Nikita Venediktov (Master's program, AI Talent Hub, ITMO; in the project — together with the Rabota.ru team)

A talk on how to build, with about one hour of algorithm runtime (and about two days of preparation), a dataset of 40 thousand short job vacancy descriptions for Supervised Fine-Tuning (SFT) of a language model, without involving annotators.

Continue reading — it's free

Google or a one-time email code. No password.