llmcmp-distill-llama3-8b-lora-v5c-target-hard-tie-response-focus-merged

Overview

Base Model

• unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit

Training Data

• Distilled source: safe_plus_filtered_plus_hard_plus_raw_partial
• Raw Kaggle train mix mode: partial
• Raw train rows are limited to non-distilled ids only
• Teacher hard mix: True

Training Format

• Prompt contains Prompt / Response A / Response B
• Teacher rationale is not injected into prompt-side input
• Completion is winner-only: A / B / C
• C means tie

Related Adapter

• Adapter repo: tussiiiii/llmcmp-distill-llama3-8b-lora-v5c-target-hard-tie-response-focus-adapter

Notes

• This model is intended for direct next-token winner inference in evaluation / submission.