Study isolates what improves learned unmasking in diffusion language models
An 856K-parameter policy on LLaDA-8B-Instruct leads the accuracy–NFE frontier on GSM8K and largely transfers to Dream-7B-Instruct.
Stories per day · 30 days
2 RSS
9 stories, newest first · one card per story
Stories tagged #decoding
Disentangling Paradigm, Identifier, and Decoding in Generative Retrieval
Decoding choice, not just the paradigm, shifts diffusion generative retrieval Hit@1 by up to 13.7 points.
A Safe Action Is Not Enough: Feasible-Future Decoding for Vision-Language-Action Policies
VICS-G, a training-free reranker, lowers frozen VLA safety cost without retraining or online rollouts.
Decoding Looped Transformers Better for (Almost) Free
LoopCD uses earlier recurrent states for contrastive decoding, boosting looped Transformers while cutting inference FLOPs.
Your Transformer Can Hold Two Thoughts at Once: Evidence of Linear Superposition in LLMs
Transformers superpose next-token distributions from mixed text streams, and fine-tuning can restore that linearity.
FLEET: From Logits Entropy to Enhanced Trajectories in Text Generation
FLEET matches repeated sampling accuracy at 3x speed and raises LiveCodeBench Pass@32 from 59.9% to 66.2%.
CARD: Cluster-level Adaptation with Reward-guided Decoding for Personalized Text Generation
CARD personalizes LLMs with cluster LoRA adapters and decoding-time preference vectors, beating LaMP and LongLaMP baselines.
On-Demand Attention: Language Models Know When to Recall
On-Demand Attention uses a lightweight recall head to selectively invoke global attention, cutting long-context decoding costs in vLLM with minimal quality loss.
Fathom: Per-Query Read Depth for Sparse Decoding over Offloaded KV Caches
Fathom introduces per-query bit-depth scanning of offloaded 4-bit KV caches, making long-context decoding 1.67x faster on Qwen3-8B at one million tokens.