Missing Claim Topic Extraction fine-tuning code

[bendavidsteel]

Hi again!:)

I'm now looking to fine-tune my own claim topic extraction model using your exact experimental setup. I see that you're using some specific techniques from the paper:
" Moreover, we use indirect
reasoning prompting, through the use of contrapositives and
contradictions in Chain-of-Thought prompting, which has been
shown to enhance the overall accuracy of factual reasoning
by 27.33% [15]. Due to lack of space, we showcase our
complete prompt-engineering framework in our repository"

So it would be great to just have your original fine-tuning script. Don't worry if it's messy I'm happy to clean it up:)