6 papers
Single-Pass Uncertainty Heads for Claim-Level Hallucination Detection in Persian Medical Language Models
Mehrdad Ghassabi, Pedram Rostami, Hamidreza Baradaran Kashani +2
Hallucination detection is particularly important for medical language models, but repeated-sampling approaches are computationally expensive. A faster alternative is a single-pass…
Gaokerena: A Small Persian Medical Language Model Family
Mehrdad Ghassabi, Hamidreza Baradaran Kashani, Pedram Rostami +5
Persian medical NLP is limited by scarce domain-specific resources and insufficient characterization of their coverage, quality, and limitations. We present Gaokerena, a suite of P…
When Detection Does Not Guarantee Resistance: Reasoning and Poisoned Context in RAG
Mehrdad Ghassabi, Audrina Ebrahimi, Sadra Hakim +3
Retrieval-Augmented Generation (RAG) exposes large language models to knowledge-poisoning attacks, where misinformation injected into retrieved documents can influence model output…
Backtranslation Augmented Direct Preference Optimization for Neural Machine Translation
Mehrdad Ghassabi, Spehr Rajabi, Hamidreza Baradaran Kashani +3
Contemporary neural machine translation (NMT) systems are almost exclusively built by training on supervised parallel data. Despite the tremendous progress achieved, these systems…
Enhancing Reasoning Skills in Small Persian Medical Language Models Can Outperform Large-Scale Data Training
Mehrdad Ghassabi, Sadra Hakim, Hamidreza Baradaran Kashani +1
Enhancing reasoning capabilities in small language models is critical for specialized applications such as medical question answering, particularly in underrepresented languages li…
Leveraging Online Data to Enhance Medical Knowledge in a Small Persian Language Model
Mehrdad Ghassabi, Pedram Rostami, Hamidreza Baradaran Kashani +3
The rapid advancement of language models has demonstrated the potential of artificial intelligence in the healthcare industry. However, small language models struggle with speciali…