Tag: efficient LLM training
Browse our exclusive articles!
OPT-BENCH: Quality-Aware RL for NP-Hard Optimization in LLMs
Discover OPT-BENCH, a quality-aware reinforcement learning framework boosting LLMs' success in NP-hard optimization with superior solution quality.
Online Reweighting Boosts LLM Training Generalization
Discover how online data reweighting outperforms offline methods in LLM training, enhancing model generalization and efficiency with ADAPT.
LoPE Boosts LLM Reasoning by Prompt Space Perturbation
Discover how LoPE uses prompt space perturbation to overcome zero-advantage issues, enhancing reasoning in large language models.
ELAS: Efficient Low-Rank LLM Pre-Training with 2:4 Sparsity
Discover ELAS, a framework that boosts low-rank LLM pre-training efficiency using 2:4 activation sparsity, reducing memory and speeding up training.