add more evaluation perspectives in terms of models

[pfliu-nlp]

• [ ] read papers and make slides:
  • On the Opportunities and Risks of Foundation Models
  • On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?
• [ ] which evaluation perspectives can we add based on the above reading?
• [ ] which tasks can we add?

[pfliu-nlp]

BTW, add more evaluation perspectives in terms of prompt view Add them into section (overleaf): 1.3

[pfliu-nlp]

More intersting papers:

• oLMpics-on what language model pre-training captures
• Measuring and improving consistency in pretrained language models
• The Values Encoded in Machine Learning Research