3 ms·
Look into RLVR (Reinforcement Learning with Verifiable Rewards). It happens during model post-training.
by pegasus 9mo ago
Look into RLVR (Reinforcement Learning with Verifiable Rewards). It happens during model post-training.
- geraneum 9mo agoThat’s not happening in run time, when you send the prompts for the next token to be generated. The comment seems to imply a run time phenomenon.