AIMarkTechPostabout 3 hours ago
Perplexity and NVIDIA Launch JOLT Method
1 min read
Perplexity and NVIDIA researchers introduce JOLT, a joint teacher-student reinforcement learning method for language models. Tested on Qwen3 variants, JOLT matches GRPO accuracy on MATH with 6.5x fewer processed training tokens, offering a direct compute solution for teams running RL post-training.
Level
Hype check
Tap to vote and see what everyone thinks.
