samfresh4/

blind-spot-fatima-qwen3-0.6B-base

	Qwen3-0.6B-Base Blind Spots Dataset




	
	Overview

This dataset documents 10 diverse cases where the base language model Qwen/Qwen3-0.6B-Base made incorrect or misleading predictions. The goal is to identify the model's "blind spots", systematic failure modes that reveal gaps in the model's reasoning, factual knowledge, and language understanding.

Model: Qwen/Qwen3-0.6B-Base Parameters: 0.6 Billion Type: Base language model (not instruction-tuned)… See the full description on the dataset page: https://huggingface.co/datasets/samfresh4/blind-spot-fatima-qwen3-0.6B-base.