THIS MIGHT INTEREST YOU

 Dear Friends,

 

Following might interest you :

 

https://medium.com/@cerebras/cerebras-makes-it-easy-to-harness-the-predictive-power-of-gpt-j-1f954f3cc6d2

 

AND,

 

https://medium.com/syncedreview/openais-webgpt-crawls-a-text-based-web-environment-to-achieve-human-level-performance-on-long-form-7dbcabe4ee44

 

Extract :

 

They employ the Microsoft Bing Web Search API for document retrieval and unsupervised pretraining and fine-tuning on the GPT-3 large language model for high-quality synthesis.

 

They then use human feedback to directly optimize answer quality, enabling their method to achieve human-level performance on LFQA tasks.

 

The team also designed a graphical interface for their text-based web-browsing environment to enable users to provide auxiliary annotations and comparison ratings to further improve the model’s understanding of the questions.

 

They evaluated the proposed WebGPT on questions from the ELI5 (Explain Like I’m 5) subreddit, with human evaluators’ judgments based on the criteria that answers should be relevant, coherent, and supported by trustworthy references.

 

 

Regards,

 

hemen