Fix wrong truncating index of tensor in DPOTrainer's concatenated_forward()

[yanghh2000]

What does this PR do?

Fixes #2330 (issue)

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline,
  Pull Request section?
• Was this discussed/approved via a GitHub issue? Please add a link
  to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines.
• Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

[kashif]

thanks @yanghh2000 would it be possible to add a test?

[yanghh2000]

Hi, I am glad to help, but I am not sure how to add a test for this. Is there any guideline to test a PR?

[yanghh2000]

Oh, I have read the guideline in trl/CONTRIBUTING.md, and what I need to do is add a test.py and commit it under test/ dir?

[kashif]

yes in the dpo_trainer tests file

[qgallouedec]

Tbh I'm not sure it is possible to test it considering it's in a middle of the method.