Enhancing large language models: toward more reliable and equitable NLP

Adiga, Rishabh

Permalink

https://hdl.handle.net/2142/129259 Copy

Description

Title

Enhancing large language models: toward more reliable and equitable NLP

Author(s)

Adiga, Rishabh

Issue Date

2025-04-28

Director of Research (if dissertation) or Advisor (if thesis)

Chandrasekaran, Varun

Department of Study

Siebel School Comp & Data Sci

Discipline

Computer Science

Degree Granting Institution

University of Illinois Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Keyword(s)

• LLM
• Few-shot
• Attention
• Bias
• Prompting
• Fairness
• Metrics
• Localization
• Mitigation

Abstract

Recent advances in large language models (LLMs) have enabled them to achieve striking fewshot performance on complex tasks, yet they can remain sensitive to prompt configurations and prone to subtle biases. This thesis addresses two central problems: (1) selecting informative few-shot examples and (2) localizing and mitigating bias in ambiguous comparative prompts. First, we propose a complexity-based approach for selecting examples in few-shot sequence tagging tasks, aiming to align test examples with training examples based on syntactic and semantic metrics. By focusing on features like sentence similarity, length matching, and label diversity, we achieve more consistent and robust outcomes without fine-tuning or adding parameters. Second, we develop a method to localize and mitigate bias in LLMs by examining the attention layers that favor certain entities over others. We then scale attention in those identified layers to reduce skewed preferences, preserving overall model fluency while mitigating biases. Extensive evaluations confirm that this targeted attention manipulation provides a lightweight way to address fairness concerns without sacrificing downstream accuracy.

Graduation Semester

2025-05

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/129259

Copyright and License Information

© 2025 Rishabh Adiga

Owning Collections