arXiv Machine Learning
PRISM: Promoting Sampling Diversity of Large Language Models with Implicit Stochastic Modulation
Read the original on arXiv Machine Learning →The Flow has not summarised this story yet — read it at arXiv Machine Learning.
Related stories
arXiv AI
Jul 22
Large Language Models Explore by Latent Distilling
arXiv:2604. 24927v2 Announce Type: replace-cross Abstract: Generating diverse responses is crucial for test-time scaling of large language models (LLMs), yet standard stochastic sampling mostly yields surface-level lexical variation, limiting semantic exploration.
arXiv Machine Learning
Aug 11
On the Effect of Sampling Diversity in Scaling LLM Inference
arXiv:2502. 11027v5 Announce Type: replace Abstract: Large language model (LLM) scaling inference is key to unlocking greater performance, and leveraging diversity has proven an effective way to enhance it.
arXiv AI
Jul 17
Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity
arXiv:2510. 01171v4 Announce Type: replace-cross Abstract: Post-training alignment often reduces LLM diversity, leading to a phenomenon known as mode collapse.
arXiv AI
Jul 1
Learning by Surprise: Adaptive Mitigation of Model Collapse in Large Language Models
arXiv:2410. 12341v4 Announce Type: replace-cross Abstract: As AI-generated content increasingly populates the web, generative AI models are at growing risk of being trained on their own outputs, a process known as AI autophagy.
arXiv AI
Oct 12
One-Step Text Generation by Seq-Drifting
The paper introduces Seq-Drifting, a new framework that enables diffusion language models to generate an entire text sequence in a single forward pass. By applying a drift field to continuous token representations, the model learns a direct mapping from Gaussian noise to text, using attraction to guide outputs toward contextually plausible tokens and repulsion to prevent repetition and collapse. Experiments show that Seq-Drifting achieves a perplexity of 36.01 on LM1B and performs competitively in unconditional, conditional, translation, summarization, and reasoning tasks.
arXiv Computation and Language
Sep 4
Breaking the Likelihood Trap: Variance-Calibrated Modulation for Large Language Model Decoding
The paper introduces Variance‑Calibrated Modulation (VCM), a training‑free pre‑decoding technique that reshapes language model probability distributions before truncation. VCM uses two dynamic mechanisms: a Contextual Searchlight via PMI to suppress stopwords and highlight context‑relevant tokens, and an Adaptive Self‑Debiasing that applies scale‑invariant penalization based on real‑time logit standard deviation. Experiments on open‑ended generation, factual QA, and mathematical reasoning show that VCM consistently reduces the likelihood trap, improving diversity, coherence, and reasoning accuracy with minimal computational cost.