RESEARCH

LLaMA appears more biased than GPT-3

By

–

25 February 2023 6h14

LLaMA is apparently more biased than GPT-3

Information about LLaMA, open language model

By

–

25 February 2023 6h14

More info on LLaMA here: https://research.facebook.com/publications/llama-open-and-efficient-foundation-language-models/ …

Meta’s new LLaMA model paper highlights PaLM’s capabilities

By

–

25 February 2023 6h14

Meta's new AI model LLaMA came out today. The paper was a gentle reminder of how capable Google's 540-billion parameter model PaLM will be.

LLaMA omits GPT’s results from MMLU test chart due to size.

By

–

25 February 2023 6h14

This was the Massive Multitask Language Understanding test (Humanities, STEM, Social Sciences, etc) When finetuned for the test, LLaMA omitted GPT's superior results from the chart because it is not of "moderate size"

→ View original post on X — @alexandr_wang

RLHF Adoption Accelerates Across Industries for Aligned AI

By

@alexandr_wang

–

25 February 2023 2h24

Today, a senior military official asked me about reinforcement learning with human feedback (RLHF). This is awesome—the widespread adoption of RLHF across every industry and use case is critical for the safe, aligned, and useful deployment of AI and LLMs. RLHF is taking over!

→ View original post on X — @bobgourley

Why Innovation is Stalling: Michael Gibson’s Perspective

By

@bobgourley

–

25 February 2023 2h22

Why is innovation stalling? Michael Gibson, a fascinating author, investor and leader, has ideas https://
youtube.com/shorts/sjD_Hp6
Jfas?feature=share
…

→ View original post on X — @thom_wolf

10B Parameter Models Vastly Underestimated in Capacity

By

@thom_wolf

–

25 February 2023 0h14

It's increasingly clear we've severely underestimated how much you can pack in a 10B parameters model