2071

LESSWRONG
LW

2070

GPT-2: 6-Month Follow-Up — LessWrong

28

GPT-2: 6-Month Follow-Up

by [anonymous]

21st Aug 2019

1 min read

28

28

GPT-2: 6-Month Follow-Up

New Comment

1 comment, sorted by

Click to highlight new comments since: Today at 11:34 AM

[-]Dr_Manhattan6y60

Also notable: NVIDIA trained a half-order-of-magnitude larger model https://nv-adlr.github.io/MegatronLM?utm_campaign=NLP%20News&utm_medium=email&utm_source=Revue%20newsletter

Curated and popular this week

Linkpost for GPT-2 6 Month Follow-Up.

Some highlights:

700+ M parameter model is being released
Several other groups have reproduced similar models
In detecting synthesized text, "current ML-based methods only achieve low to mid–90s accuracy"