PaLM 2
Free while signed in. Answers cite the passages they came from.
Paper preview
The curator’s take
Key pointsGoogle's second-generation PaLM powering Bard and Google products.
01
Compute-optimal training: Trained compute-optimally on a larger, higher-quality, more multilingual corpus than PaLM 1.
02
Multilingual strength: Major improvement in 100+ languages; supports translation, generation, and reasoning across a much broader language set.
03
Reasoning competitive with GPT-4: Particularly strong on mathematical reasoning, approaching GPT-4 on several benchmarks.
04
Flan-PaLM 2: The instruction-tuned version performs well on MMLU, BIG-bench Hard, and code generation - powering Google's consumer AI products.
Every Monday
Get next week’s papers.
The same picks and the same summaries, in your inbox. Free, and 176 issues deep.
Subscribe on Substack