Papers
arxiv:2611.25735

PosteriorBench: Evaluating Behavior Posterior Sampling in Language Models

Published on Nov 19
Authors:
,
,
,

Abstract

PosteriorBench benchmarks algorithms for sampling rare language model behaviors, measuring compute cost and closeness to the true posterior, and reveals that no evaluated method is both efficient and faithful.

Rare language model behaviors can be difficult to study because obtaining them through ordinary sampling can require a very large number of samples. This motivates the behavior posterior sampling problem: given a language model and a behavior of interest, is there an algorithm that efficiently samples from the model's distribution over transcripts that exhibit this behavior? We introduce PosteriorBench, a benchmark for comparing algorithms by their compute cost and by how close their samples are to the true posterior in forward and reverse KL divergence. PosteriorBench contains 21 tasks on Qwen3.5-0.8B with behavior probabilities from 10^{-5} to 10^{-1}, each with exact posterior samples from a corpus of one billion labeled responses. We find that no method we evaluate is both efficient and faithful: simple methods like prompting are cheap but plateau at a high KL divergence from the posterior, fine-tuning on posterior samples is often cost-prohibitive, and several training methods from prior work do not perform better. Building a posterior sampler that can cheaply approximate the true posterior remains an open problem, and we hope PosteriorBench can measure progress towards this objective. Code and data are available at https://github.com/TransluceAI/PosteriorBench.

Community

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2611.25735
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2611.25735 in a model README.md to link it from this page.

Datasets citing this paper 2

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2611.25735 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.