Evaluating Language Models
Walid and Gábor discuss the challenges of evaluating language models, emphasizing the limitations of using perplexity as a sole metric. Gábor highlights the importance of optimizing for human perception rather than relying solely on proxies like perplexity or machine translation metrics.In this clip
From this podcast
NLP Highlights
40 - On the State of the Art of Evaluation in Neural Language Models, with Gábor Melis
Related Questions