paper-with-me

Papers

Language Model Maps for Prompt-Response Distributions via Log-Likelihood Vectors

2026-03-19 · Yusuke Takase, Momose Oyama, Hidetoshi Shimodaira arxiv

We propose a method that represents language models by log-likelihood vectors over prompt-response pairs and constructs model maps for comparing their conditional distributions. In this space, distances between models approximate the KL divergence between the corresponding conditional distributions. Experiments on a large collection of publicly available language models show that the maps capture meaningful global structure, including relationships to model attributes and task performance. The method also captures systematic shifts induced by prompt modifications and their approximate additive compositionality, suggesting a way to analyze and predict the effects of composite prompt operations. We further introduce pointwise mutual information (PMI) vectors to reduce the influence of unconditional distributions; in some cases, PMI-based model maps better reflect training-data-related differences. Overall, the framework supports the analysis of input-dependent model behavior.

📄 PDF Abstract BibTeX arXiv:2603.18593

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Deep Convolutional Likelihood Particle Filter for Visual Tracking

2020-06-11 · Reza Jalil Mozhdehi, Henry Medeiros

We propose a novel particle filter for convolutional-correlation visual trackers. Our method uses correlation response maps to estimate likelihood distributions and employs these likelihoods as proposal densities to samp…

Visual Tracking

CoRT: Counterfactual Replay for Token-Level Rubric-Guided Policy Optimization

2026-07-28 · Bo-Wen Zhang, Junwei He, Wen Wang, Song-Lin Lv 외 arxiv

Rubric-based reinforcement learning enriches language model training by evaluating model outputs against explicit criteria. Yet in GRPO-style pipelines, these structured judgments are reduced to a scalar response-level r…

Reinforcement Learning

Quantitative Certification of Bias in Large Language Models

2024-05-29 · Isha Chaudhary, Qian Hu, Manoj Kumar, Morteza Ziyadi 외

Large Language Models (LLMs) can produce biased responses that can cause representational harms. However, conventional studies are insufficient to thoroughly evaluate LLM bias, as they can not scale to large number of in…

Benchmarking

Likelihood as a Performance Gauge for Retrieval-Augmented Generation

2024-11-12 · Tianyu Liu, Jirui Qi, Paul He, Arianna Bisazza 외

Recent work finds that retrieval-augmented generation with large language models is prone to be influenced by the order of retrieved documents in the context. However, the lack of in-depth analysis limits the use of this…

Language ModelingLanguage ModellingPrompt EngineeringQuestion Answering+2

Differential Privacy in Generative AI Agents: Analysis and Optimal Tradeoffs

2026-03-18 · Ya-Ting Yang, Quanyan Zhu arxiv

Large language models (LLMs) and AI agents are increasingly integrated into enterprise systems to access internal databases and generate context-aware responses. While such integration improves productivity and decision …

Response Generation