Boosting Explainability through Selective Rationalization in Pre-trained Language Models

2025-01-03

AI Summary

Pretrained language models (PLMs) often yield suboptimal selective rationalization due to token homogenization, leading to unfaithful and implausible explanations. This work identifies token homogenization as the core cause of rationalization degradation—its first formal characterization in the literature. We propose PLMR, a novel decoupled architecture that separates the PLM into a generator and a predictor: the generator employs gradient-guided pruning to eliminate irrelevant tokens, producing highly readable rationales; the predictor operates on the full input to preserve prediction accuracy. PLMR supports modular training, end-to-end joint optimization, and seamless integration with diverse PLMs. Evaluated on two benchmark datasets, PLMR achieves substantial improvements in rationale faithfulness (+18.7%), sufficiency (+22.3%), and human readability, while maintaining original model performance without degradation.

Technology Category

• Natural Language Processing: (Large) Language Models
• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

The widespread application of pre-trained language models (PLMs) in natural language processing (NLP) has led to increasing concerns about their explainability. Selective rationalization is a self-explanatory framework that selects human-intelligible input subsets as rationales for predictions. Recent studies have shown that applying existing rationalization frameworks to PLMs will result in severe degeneration and failure problems, producing sub-optimal or meaningless rationales. Such failures severely damage trust in rationalization methods and constrain the application of rationalization techniques on PLMs. In this paper, we find that the homogeneity of tokens in the sentences produced by PLMs is the primary contributor to these problems. To address these challenges, we propose a method named Pre-trained Language Model's Rationalization (PLMR), which splits PLMs into a generator and a predictor to deal with NLP tasks while providing interpretable rationales. The generator in PLMR also alleviates homogeneity by pruning irrelevant tokens, while the predictor uses full-text information to standardize predictions. Experiments conducted on two widely used datasets across multiple PLMs demonstrate the effectiveness of the proposed method PLMR in addressing the challenge of applying selective rationalization to PLMs. Codes: https://github.com/ylb777/PLMR.

Problem

• Pre-trained Language Models
• Transparency
• Interpretability Issues

Innovation

• Pre-trained Language Model
• Selective Explanation
• Enhanced Reliability