bbhqwen6 / README.md
sumink's picture
Create README.md
1b75baa verified
๐Ÿš€ Qwen2.5-3B Fine-Tuned on BBH (Hyperbaton) - Model Card
๐Ÿ“Œ Model Overview
Model Name: Qwen2.5-3B Fine-Tuned on BBH (Hyperbaton)
Base Model: Qwen2.5-3B-Instruct
Fine-Tuned Dataset: BBH (BigBench Hard) - Hyperbaton
Task: Adjective Ordering Classification
Fine-Tuning Objective: Train the model to select the correct English adjective order in a sentence.
๐Ÿ“Œ Dataset Information
This model was fine-tuned on the Hyperbaton subset of the BigBench Hard (BBH) dataset.
Dataset characteristics:
Task Type: Correct adjective order classification
Input Format: Two sentence options where one follows the correct adjective order
Target Labels: "A" (correct) or "B" (incorrect)
Example:
Input:
"Which sentence has the correct adjective order? Options: (A) midsize old grey Brazilian sweater (B) midsize grey Brazilian old sweater"
Target: "A"
This dataset evaluates a modelโ€™s ability to understand syntactic correctness in English adjective order, which is useful for natural language generation, grammar checking, and language learning applications.