REVIEW 5 cited by
TokenSHAP: Interpreting Large Language Models with Monte Carlo Shapley Value Estimation
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
As large language models (LLMs) become increasingly prevalent in critical applications, the need for interpretable AI has grown. We introduce TokenSHAP, a novel method for interpreting LLMs by attributing importance to individual tokens or substrings within input prompts. This approach adapts Shapley values from cooperative game theory to natural language processing, offering a rigorous framework for understanding how different parts of an input contribute to a model's response. TokenSHAP leverages Monte Carlo sampling for computational efficiency, providing interpretable, quantitative measures of token importance. We demonstrate its efficacy across diverse prompts and LLM architectures, showing consistent improvements over existing baselines in alignment with human judgments, faithfulness to model behavior, and consistency. Our method's ability to capture nuanced interactions between tokens provides valuable insights into LLM behavior, enhancing model transparency, improving prompt engineering, and aiding in the development of more reliable AI systems. TokenSHAP represents a significant step towards the necessary interpretability for responsible AI deployment, contributing to the broader goal of creating more transparent, accountable, and trustworthy AI systems.
Forward citations
Cited by 5 Pith papers
-
Shapley Context Pruning: A Cooperative Game Perspective for Context Reranking and Pruning
SCP ranks and prunes context sentences using Shapley values from a learned Deep Sets value function, matching or beating baselines on several multi-hop QA datasets at 50% compression.
-
PLEX: Perturbation-free Local Explanations for LLM-Based Text Classification
PLEX learns a mapping from BERT or RoBERTa token embeddings to word importance scores, reproducing LIME and SHAP style explanations without per-sentence perturbations.
-
Demystifying Reasoning Dynamics with Mutual Information: Thinking Tokens are Information Peaks in LLM Reasoning
Reasoning tokens like 'Hmm' and 'Wait' mark steps where a model's internal state carries unusually high dependence with the correct answer, and suppressing them hurts accuracy.
-
Model-Agnostic FDR Control via Group Gaussian Mirror and Permutation SHAP
Block-level Gaussian mirror statistics give a mostly sound linear FDR method, but the neural Permutation SHAP variant proves null symmetry only by assuming the fitted model already ignores null groups.
-
SCAR: Shapley Credit Assignment for More Efficient RLHF
SCAR redistributes the terminal RLHF reward to tokens and spans via Shapley values, preserving the total return while improving training efficiency and final reward across three LLM alignment tasks.
Discussion (0). Continue with ORCID to comment.