Instruction-Tuned Llama-3.2-3B LLM (Data Labeling & Fine-tuning)
OtherTextFine Tuning
For the Instruction-Tuned Llama-3.2-3B LLM project, I engineered and labeled a dataset for supervised fine-tuning. The process entailed curating instruction-response pairs and employing prompt engineering techniques to improve model performance. My work played a key role in preparing high-quality data for LLM fine-tuning and evaluation.
• Constructed and labeled approximately 12,000 instruction-response text samples.
• Applied chain-of-thought prompt strategies to enhance dataset effectiveness.
• Tuned model inference parameters based on labeled training outcomes.
• Conducted quality assurance by reviewing and annotating dataset entries.
For the Instruction-Tuned Llama-3.2-3B LLM project, I engineered and labeled a dataset for supervised fine-tuning. The process entailed curating instruction-response pairs and employing prompt engineering techniques to improve model performance. My work played a key role in preparing high-quality data for LLM fine-tuning and evaluation. • Constructed and labeled approximately 12,000 instruction-response text samples. • Applied chain-of-thought prompt strategies to enhance dataset effectiveness. • Tuned model inference parameters based on labeled training outcomes. • Conducted quality assurance by reviewing and annotating dataset entries.
Not specified