Fatimoh/

qwen2.5-1.5b-blind-spots

	Qwen2.5-1.5B Blind Spots Dataset




	
	Overview

This dataset was created as part of the Fatima Fellowship 2026 Technical Challenge. It documents cases where Qwen/Qwen2.5-1.5B a 1.5 billion parameter base language model produces incorrect or unexpected outputs across 10 diverse input categories.

	Model Tested




	

Field Detail

Model Qwen/Qwen2.5-1.5B

Parameters ~1.5 billion

Type Base language model (not instruction-tuned)

Modality Text (LLM)

Model Card… See the full description on the dataset page: https://huggingface.co/datasets/Fatimoh/qwen2.5-1.5b-blind-spots.