Template collapse is a distinct failure mode in agentic RL invisible to entropy; mutual information proxies diagnose it better and SNR-aware filtering using reward variance improves input-dependent reasoning and task performance across planning, math, navigation, and code tasks.
Dapo: An open-source llm reinforcement learning system at scale
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
citation-role summary
dataset 1
citation-polarity summary
verdicts
UNVERDICTED 4roles
dataset 1polarities
use dataset 1representative citing papers
SPaCe uses semantic clustering to shrink training sets and a multi-armed bandit to adaptively select samples, matching or beating baselines on reasoning benchmarks with up to 100x fewer examples.
On-policy self-distillation fails for instance-specific privileged information because the student learns an aggregated PI-free policy, while on-policy distillation is sensitive to teacher choice and loss formulation, with stop-gradient and stabilized methods as mitigations.
HAPO is a new token-level policy optimization method for LLMs that continuously adapts four optimization stages using entropy, claiming consistent gains over DAPO on math, code, and logic tasks.
citing papers explorer
-
SPaCe: Unlocking Sample-Efficient Large Language Models Training With Self-Pace Curriculum Learning
SPaCe uses semantic clustering to shrink training sets and a multi-armed bandit to adaptively select samples, matching or beating baselines on reasoning benchmarks with up to 100x fewer examples.
-
Heterogeneous Adaptive Policy Optimization: Tailoring Optimization to Every Token's Nature
HAPO is a new token-level policy optimization method for LLMs that continuously adapts four optimization stages using entropy, claiming consistent gains over DAPO on math, code, and logic tasks.