BehaviorBox: Automated Discovery of Fine-Grained Performance Differences Between Language Models

TL;DR

BehaviorBox is an automated methodology that identifies fine-grained, context-specific performance differences between language models by extracting meaningful features where one model outperforms another, aiding deeper understanding beyond traditional metrics.

Contribution

This work introduces BehaviorBox, a novel automated approach for discovering detailed performance differences between language models using performance-aware contextual embeddings.

Findings

Identifies specific contextual features where models differ in performance

Reveals insights not captured by corpus-level perplexity measures

Applies to various models, sizes, and training methods

Abstract

Language model evaluation is a daunting task: prompts are brittle, corpus-level perplexities are vague, and the choice of benchmarks are endless. Finding examples that show meaningful, generalizable differences between two LMs is crucial to understanding where one model succeeds and another fails. Can this process be done automatically? In this work, we propose methodology for automated comparison of language models that uses performance-aware contextual embeddings to find fine-grained features of text where one LM outperforms another. Our method, which we name BehaviorBox, extracts coherent features that demonstrate differences with respect to the ease of generation between two LMs. Specifically, BehaviorBox finds features that describe groups of words in fine-grained contexts, such as "conditional 'were' in the phrase 'if you were'" and "exclamation marks after emotional statements", where one model outperforms another within a particular datatset. We apply BehaviorBox to compare models that vary in size, model family, and post-training, and enumerate insights into specific contexts that illustrate meaningful differences in performance which cannot be found by measures such as corpus-level perplexity alone.