What is post-training?
Post-training is the set of steps after pretraining, such as supervised fine-tuning and reinforcement learning, that shape a model's behavior.
Last reviewed
post-training in AI data
It relies on smaller, higher-quality datasets, often written or judged by humans, which is why expert data commands high prices.
Type: training stage.
Read more: guides / rlhf vs sft.
Where post-training applies on this site
- Data type: Supervised fine-tuning data
- Guide: Who buys AI training data?
- Guide: RLHF vs SFT: what is the difference?
Quick answers
What is post-training?
Post-training is the set of steps after pretraining, such as supervised fine-tuning and reinforcement learning, that shape a model's behavior.
Why does post-training matter in AI data?
It relies on smaller, higher-quality datasets, often written or judged by humans, which is why expert data commands high prices.