As a user, I want MT metric scores for the machine translations of the Notion parallel sentence corpus, so I can evaluate translations of different models.

Need to generate the following 4 CSVs:

- [ ] Google's TLLM
- [ ] HYT-MT1.5
- [ ] NLLB-200
- [ ] MADLAD-400

Depends on #1, #16 