871 private links
A really innovative way to store exact floating point numbers:
- for "safe numbers" (numbers whose value can be expressed precisely in floating point), store them as is
- for non-"safe numbers", store two safe numbers as denominator and nominator.
There is a lot of this method comparing to the other alternatives:
- BigDecimal: a lot faster than arbitrary precision BigDecimal
- Rational: denom and nom deoesn't blow up to extremely large numbers
And it is also faster and uses less space.
Nice visualization of dynamic math. Attractors, vector fields, fractal, flocking, and more.
An actually informative guide on prompt engineering. It talks about how to communicate format, prompt injection, how to prompt hack, etc.
What are the simple questions to ask that exploit the limitation of LLM? Interesting examples.
Disposable email service. Similar to temp-mail.org but doesn't have the Cloudflare traffic check, so it works even with VPN on.
De-DRM audible audio books.
Alternative URLs:
A small webpage tool to play with instrumental music. Synthesizer, composer, sequencer.
Here's a more powerful one: https://hidden-realm.github.io/cardboardbox, it has additional features like customizing harmonics.
Tools and methods for investigating linux disk I/O.
LSTM gate mechanism clearly explained. I read this article years ago but it's nice to read it again.
This article discusses conventional and newer fine tuning techniques for LLM. Keyword for search: prompt tuning, prefix tuning, in-context learning, frozen layers, adaptor, LoRA.
Explore AI generated arts from various models.
An article about color vision, color space, gamma.
LLM prompting tricks and patterns.
A catalogue of AI tools and news.
A post on fine-tuning Alpaca with a lot of workable actions.
This article talks about the evolution of HTTP versions. From the earliest HTTP/0.9 where there is just a single bare GET method and supports no headers, to HTTP/3 which is completely optimized for speed and security. By the way, HTTP/0.9 resembles FTP a lot, which is interesting.
A nice showcase of powershell features.
Visualize tokenization for OpenAI models.
“Quantity creates emergence. Simple elements, complex interactions, new patterns.” -- by Bing Chat
I am always fascinated by how interesting phenomenon seem to emerge suddenly in complex systems. The recent breakthrough in Large Language Models has brought me surprises after surprises.
This article aggregates a list of articles on emergent capabilities found in LLM.
This site collects rare and (mostly) old movies not easily found on the internet.