HackerTrans
TopNewTrendsCommentsPastAskShowJobs

Homunculiheaded

no profile record

Submissions

[untitled]

1 points·by Homunculiheaded·2 jaar geleden·0 comments

[untitled]

1 points·by Homunculiheaded·2 jaar geleden·0 comments

Making Structured Generation Faster Than Unstructured

blog.dottxt.co
4 points·by Homunculiheaded·2 jaar geleden·0 comments

Coding for Structured Generation with LLMs

blog.dottxt.co
4 points·by Homunculiheaded·2 jaar geleden·0 comments

Beating GPT-4 with Open Source

blog.dottxt.co
5 points·by Homunculiheaded·2 jaar geleden·0 comments

Prompt Efficiency – Getting 8-shot LLM performance from 1-shot

blog.dottxt.co
2 points·by Homunculiheaded·2 jaar geleden·0 comments

How fast can CFG-Structured generation be?

blog.dottxt.co
7 points·by Homunculiheaded·2 jaar geleden·0 comments

Structured Generation Improves LLM Performance: GSM8K Benchmark

blog.dottxt.co
11 points·by Homunculiheaded·2 jaar geleden·4 comments

Coalescence: Making LLM inference 5x faster

blog.dottxt.co
6 points·by Homunculiheaded·2 jaar geleden·0 comments

Linear Diffusion

github.com
2 points·by Homunculiheaded·3 jaar geleden·0 comments

Linear Diffusion: Building a Diffusion Model from Linear Components

countbayesie.com
3 points·by Homunculiheaded·3 jaar geleden·0 comments

Replacing an A/B Test with GPT

countbayesie.com
3 points·by Homunculiheaded·3 jaar geleden·0 comments

comments

Homunculiheaded
·2 jaar geleden·discuss
Agreed! While these results are very promising, there's still a lot to explore in this space.

In addition to the "prompt consistency" and "thought-control" ideas mentioned in the post, I'm definitely curious how the performance is on more complex structured data (things like codegen).