NI Sampling accelerates discrete diffusion language models up to 14.3 times by training a neural indicator to select which tokens to sample at each step using a trajectory-preserving objective.
R2r: Efficiently navigating divergent reasoning paths with small-large model token routing.arXiv preprint arXiv:2505.21600
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
years
2026 5representative citing papers
SOL trains a policy to dynamically control multiple efficiency mechanisms per token via group-relative policy optimization on teacher-forced episodes, yielding better quality at matched average budget than static or random allocation.
Small language models can achieve near large-model reasoning performance by learning to re-rank their own top-K token predictions after distilling selection from the large model.
GlimpRouter uses the entropy of the first token in each reasoning step to decide whether to invoke a large model, yielding 10.7% higher accuracy and 25.9% lower latency than a standalone large model on AIME25.
FusionRoute augments token-level expert routing with a trainable complementary logit generator to expand the policy class and recover optimal decoding under mild conditions, outperforming prior collaboration and merging methods on reasoning and generation benchmarks.
citing papers explorer
-
NI Sampling: Accelerating Discrete Diffusion Sampling by Token Order Optimization
NI Sampling accelerates discrete diffusion language models up to 14.3 times by training a neural indicator to select which tokens to sample at each step using a trajectory-preserving objective.
-
Compute Where it Counts: Self Optimizing Language Models
SOL trains a policy to dynamically control multiple efficiency mechanisms per token via group-relative policy optimization on teacher-forced episodes, yielding better quality at matched average budget than static or random allocation.
-
Select to Think: Unlocking SLM Potential with Local Sufficiency
Small language models can achieve near large-model reasoning performance by learning to re-rank their own top-K token predictions after distilling selection from the large model.
-
GlimpRouter: Efficient Collaborative Inference by Glimpsing One Token of Thoughts
GlimpRouter uses the entropy of the first token in each reasoning step to decide whether to invoke a large model, yielding 10.7% higher accuracy and 25.9% lower latency than a standalone large model on AIME25.
-
Token-Level LLM Collaboration via FusionRoute
FusionRoute augments token-level expert routing with a trainable complementary logit generator to expand the policy class and recover optimal decoding under mild conditions, outperforming prior collaboration and merging methods on reasoning and generation benchmarks.