ICML accepts prompt-engineering paper 'Verbalized Sampling' sparking debate on rigor
A paper titled 'Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity' has been accepted to ICML 2026. The work proposes a simple prompt-engineering trick to improve sampling diversity, but its acceptance at a top-tier ML conference has drawn criticism for lacking rigorous theoretical analysis.
Entities
Related
Users report 'show me three versions' prompt trick improves output diversity
A Reddit user shared a prompting technique that asks for three distinct versions of an answer at once, rather than iterating one at a time. The method reportedly beats the regression-to-the-average problem by forcing the model to produce genuinely different outputs. The post has garnered 11 engagement points.
NL-PAC paper introduces framework for handling specification ambiguity in LLM supervision
A new arXiv paper proposes NL-PAC, a framework that addresses specification ambiguity when LLMs provide labels or evaluations from natural language instructions. The framework uses a fixed model's thresholded decoding law to define admissible labels and candidate targets, showing that additional labels reduce sampling error but cannot resolve identification problems from ambiguous specifications.
LLM/VLA models enable prompt-driven exploration in RL
A new research paper proposes using large language models (LLMs) and vision-language-action (VLA) models to drive exploration in reinforcement learning by modifying natural language prompts, which induce global behavioral changes beyond standard action noise. This approach could help policies escape weak local optima more effectively.
Researchers identify asymmetric generalization problem in LLM unlearning benchmarks
A new arXiv paper argues that existing machine unlearning benchmarks for LLMs suffer from under-forgetting and over-forgetting due to an asymmetric generalization problem. The authors propose that evaluation must cover diverse query formulations of target facts to reliably measure knowledge removal while preserving unrelated capabilities.
Researchers identify modality interference as root cause of full-duplex SLM degradation
A new paper on arXiv (July 7, 2026) presents a fine-grained analysis of optimization dynamics in full-duplex Spoken Language Models (SLMs), identifying severe modality interference as the root cause of knowledge degradation and compromised semantic integrity. The work aims to enable more natural and intelligent full-duplex SLMs.