What question did this study set out to answer?

The aim is to analyze explainability techniques for large language models in order to improve interpretability and trust in AI decisions.

March 21, 2026Open Access

Applied Explainability for Large Language Models: A Comparative Study

View Full Paper

Ask AI

Bookmark

Authors

Comparative analysis evaluates explainability techniques in sentiment classification, highlighting practical trade-offs.

The aim is to analyze explainability techniques for large language models in order to improve interpretability and trust in AI decisions.
Applied three explainability techniques: Integrated Gradients, Attention Rollout, and SHAP.
Utilized a fine-tuned DistilBERT model for SST-2 sentiment classification.
Evaluated methods based on qualitative criteria such as faithfulness, stability, and interpretability.
Gradient-based methods provided the most stable and intuitive explanations.
Attention-based approaches are computationally efficient but less aligned with prediction-relevant features.
Model-agnostic methods introduced flexibility but came with computational overhead and variability.

Venkata Abhinandan Kancharla (2026) studied this question.

View Full Paper

Ask AI

Bookmark

March 21, 2026Open Access

View Full Paper

Ask AI

Bookmark

Authors

Comparative analysis evaluates explainability techniques in sentiment classification, highlighting practical trade-offs.

The aim is to analyze explainability techniques for large language models in order to improve interpretability and trust in AI decisions.
Applied three explainability techniques: Integrated Gradients, Attention Rollout, and SHAP.
Utilized a fine-tuned DistilBERT model for SST-2 sentiment classification.
Evaluated methods based on qualitative criteria such as faithfulness, stability, and interpretability.
Gradient-based methods provided the most stable and intuitive explanations.
Attention-based approaches are computationally efficient but less aligned with prediction-relevant features.
Model-agnostic methods introduced flexibility but came with computational overhead and variability.

Venkata Abhinandan Kancharla (2026) studied this question.

View Full Paper

Ask AI

Bookmark