Towards Trustworthy Large Language Models

Sakshi Parate; Shreyans Sanyal

Towards Trustworthy Large Language Models

Author(s)	Ms. Sakshi Parate, Mr. Shreyans Sanyal
Country	India
Abstract	Large language models (LLMs) are increasingly deployed in domains where opacity, factual un-reliability, and computational cost carry real consequences, yet the research communities that address these three problems — explainability, hallucination, and inference efficiency — have largely developed in isolation. This paper argues that the three concerns are not independent: interventions designed to reduce inference cost, such as prompt compression, can silently alter both the faithfulness of post-hoc explanations and the model’s propensity to hallucinate. We present an integrated conceptual frame-work that couples attention- and perturbation-based explainability (LIME, SHAP, raw attention) with lightweight hallucination-detection signals and token-efficient inference strategies (chunking, summa-rization, prompt compression), and we instrument the framework with a set of cross-cutting consistency metrics. To ground the framework empirically, we design and execute a controlled pilot experiment that measures how rule-based prompt compression jointly affects (i) the Kullback–Leibler divergence between a model’s output distributions before and after compression, used as a proxy for hallucination risk, and (ii) the entropy and rank-correlation of last-layer attention, used as a proxy for explanation faithfulness. Because the sandboxed experimental environment used for this study has no network ac-cess to pretrained model repositories, the experiment is conducted on a compact two-layer Transformer language model trained from scratch on a controlled synthetic corpus, which allows exact, reproducible control over ground truth while preserving the qualitative mechanics of attention-based attribution and next-token prediction under compression. Across 15 held-out prompts, compression removes 52.9% of tokens on average while producing a small but non-zero mean output KL divergence of 0.0007 and re-ducing mean attention entropy from 1.062 to 0.632, with attention rank correlation across compression conditions of only 0.633. These results provide direct, quantitative evidence for the framework’s central hypothesis: token-efficiency interventions are not explanation-neutral, and systems that report compres-sion ratios without also reporting faithfulness and hallucination-risk deltas may be masking a three-way trade-off. We discuss the implications for trustworthy LLM system design and outline how the pilot findings motivate follow-up experiments on larger pretrained models.
Keywords	LLM, Hallucination, Explanability, Token Efficiency
Field	Computer > Artificial Intelligence / Simulation / Virtual Reality
Published In	Volume 7, Issue 4, July-August 2026
Published On	2026-08-15

View / Download PDF File

E-ISSN 3048-7641

doi

CrossRef DOI is assigned to each research paper published in our journal.

AIJFR DOI prefix is
10.63363/aijfr

Downloads

Research Paper Format Copyright Permission Form and Undertaking Form Cover Page Vol 7 Isu 4 Cover Page Vol 7 Isu 3 Cover Page Vol 7 Isu 2

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA

About AIJFR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9898-79-39-59	editor@aijfr.com

Advanced International Journal for Research

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Towards Trustworthy Large Language Models

Share this