all AI news
Quoting PaLM 2 Technical Report (PDF)
Simon Willison's Weblog simonwillison.net
The largest model in the PaLM 2 family, PaLM 2-L, is significantly smaller than the largest PaLM model but uses more training compute. Our evaluation results show that PaLM 2 models significantly outperform PaLM on a variety of tasks, including natural language generation, translation, and reasoning. These results suggest that model scaling is not the only way to improve performance. Instead, performance can be unlocked by meticulous data selection and efficient architecture/objectives. Moreover, a smaller but higher quality model significantly …
ai bard compute evaluation family generativeai google language language generation llms natural natural language natural language generation palm palm 2 pdf reasoning report scaling show technical training translation