Learning Adaptive Parallel Reasoning with Language Models Paper • 2504.15466 • Published 6 days ago • 39
view post Post 393 We create a dataset of 1 million, MIT-licensed synthetic humans, sampled from actual US demographics. You can use it to seed LLM synthetic data generation and create extremely diverse, statistically realistic outputs. Dataset: skysight-inc/synthetic-humans-1m Accompanying blog post with methodology: https://www.skysight.inc/blog/synthetic-humans See translation 👍 1 1 + Reply
LMRL Gym: Benchmarks for Multi-Turn Reinforcement Learning with Language Models Paper • 2311.18232 • Published Nov 30, 2023 • 1
Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters Paper • 2408.03314 • Published Aug 6, 2024 • 63