S M Tahmid Siddiqui
1 indexed paper
Recent (6 mo)
1With code
0Influential cites
0Benchmarked
0Publications per year
126
Top categories
NLP×1ML×1
Frequent co-authors
Research Timeline
2026
Towards Lightweight Reliability: Using Soft Prompts for Hallucination Mitigation in Large Language Models
The paper introduces Responsible Contrastive Soft Prompting (RCSP), a parameter-efficient method using soft prompts to improve LLM reliability by simultaneously suppressing hallucinations, encouraging abstention, and maintaining factual recall.
Highlighted terms show continued research focus across papers