3 ms·
"The human evaluation results indicate that the response generated from DialoGPT is comparable to human response quality under a single-turn conversation Turing
by vok 7y ago
"The human evaluation results indicate that the response generated from DialoGPT is comparable to human response quality under a single-turn conversation Turing test."
Paper here:
https://arxiv.org/abs/1911.00536 https://arxiv.org/abs/1911.00536