Terminal-World is a skill-based synthesis pipeline that generates 5,723 training environments and produces Terminal-World-32B which outperforms baselines on Terminal-Bench 2.0 using only 1.2% of the data.
American invitational mathematics examination (aime) 2025
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
citation-role summary
dataset 1
citation-polarity summary
years
2026 4verdicts
UNVERDICTED 4roles
dataset 1polarities
use dataset 1representative citing papers
A new RL objective adapts trust-region and off-policy handling automatically via normalized effective sample size of batch policy ratios, matching tuned baselines without new hyperparameters.
GSQ uses Gumbel-Softmax to optimize scalar quantization grids for LLMs, closing most of the accuracy gap to vector methods like QTIP at 2-3 bits per parameter while using symmetric scalar grids compatible with existing kernels.
RankGuide uses tensor-rank analysis of consecutive hidden states to route between small and large reasoning models and steer generations, reducing latency up to 1.75x while maintaining competitive accuracy on reasoning benchmarks.
citing papers explorer
-
Terminal-World: Scaling Terminal-Agent Environments via Agent Skills
Terminal-World is a skill-based synthesis pipeline that generates 5,723 training environments and produces Terminal-World-32B which outperforms baselines on Terminal-Bench 2.0 using only 1.2% of the data.
-
Trust the Batch, On- or Off-Policy: Adaptive Policy Optimization for RL Post-Training
A new RL objective adapts trust-region and off-policy handling automatically via normalized effective sample size of batch policy ratios, matching tuned baselines without new hyperparameters.
-
GSQ: Highly-Accurate Low-Precision Scalar Quantization for LLMs via Gumbel-Softmax Sampling
GSQ uses Gumbel-Softmax to optimize scalar quantization grids for LLMs, closing most of the accuracy gap to vector methods like QTIP at 2-3 bits per parameter while using symmetric scalar grids compatible with existing kernels.
-
RankGuide: Tensor-Rank-Guided Routing and Steering for Efficient Reasoning
RankGuide uses tensor-rank analysis of consecutive hidden states to route between small and large reasoning models and steer generations, reducing latency up to 1.75x while maintaining competitive accuracy on reasoning benchmarks.