Zephyr-dpo-full Checkpoints perform poorly on TruthfulQA.

Hello,
I observe that the models I trained and the official models provided by [HuggingFace](https://huggingface.co/alignment-handbook/) do not match the results of [Zephyr-7b-beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta) on TruthfulQA.

I used lm_evaluate_harness for evaluation, and the metric used was mc2.

The result for HuggingFaceH4/zephyr-7b-beta is 55.15, and the result for mistralai/Mistral-7B-7b-beta is 42.59. Both of these numbers are correct.

However, the result for alignment-handbook/zephyr-7b-dpo-full is only 45.07, and the results for alignment-handbook/zephyr-7b-sft-full is only 40.38.

Furthermore, the results from my own trained checkpoint are also incorrect. The result for sft-full is 40.12, and the result for dpo-full is 47.40.

Version of lm-evaluate-harness is [this](https://github.com/EleutherAI/lm-evaluation-harness/commit/994bdb3fd3780110e45f479f5f409cae05e089bd)