#natural-language-inference

2 questions from NLP Tasks

Timed test

2 questions · 2 min · score and explanations at the end

Q 1 / 2Score 0 / 0

Shuffle questions

Question 1: Single choice

mediumAI-checkedNLP Tasks

How could a natural language inference (NLI) model judge a machine translation against a reference translation, without relying on n-gram overlap?

Check entailment both ways: output and reference that entail each other mean the same thing1
Check only that the reference entails the output, since one direction suffices2
Reward translations that the NLI model labels "neutral" with respect to the reference3
Score how strongly the translation contradicts the source sentence and prefer low values4

Keys: 1–9 select · Enter check / next