zwaps

11 ms·

151.
by zwaps 2y ago
We do not, and they don’t. Almost every employee starts out believing the spiel about mission, team, and caring for workers (to some degree) and personally invest care in the endeavor. Up until they are fired for the first time seemingly in
152.
by zwaps 2y ago
Why is it that all of the many eval and LLM ops offerings spent seemingly all their energy on UI and playgrounds? When it comes to tracking, tracing and versioning the entire LLM callchain, so from prompt, to response models, model and work
153.
by zwaps 2y ago
This is called speculative decoding
154.
by zwaps 2y ago
(German) offshore seems to be missing
155.
by zwaps 2y ago
What is the difference to HippoRAG, which seems to be the same approach but came our earlier?
156.
by zwaps 2y ago
Different field entirely
157.
by zwaps 2y ago
Am i blind? Where is the code?
158.
by zwaps 2y ago
This says too little of substance. What errors were made in YCs assessment, which will be corrected? What is the path forward?
159.
by zwaps 2y ago
The site insinuates that it is a set of flexible and probably parametric mix between full attention and linear conplexity models
160.
by zwaps 2y ago
It sounds like you are training multiple low rank adapters?
161.
by zwaps 2y ago
Exactly, because Glove embeddings generate directly a linear space. This is, if I remember the paper, explicitly their objective. This corresponds to a single symmetrical relationship between two tokens, so a rather simple network embeddabl
162.
by zwaps 2y ago
Like you said, language models have started "working" to write text since (and not before) they used attention, that is, adjacency matrices (being GNNs). GloVe and Word2Vec did well for aggregate word representations over an entir
163.
by zwaps 2y ago
If I click on this page I get something about Wedding AI. Joke?
164.
by zwaps 2y ago
Yeah but in the LLM/AI space, everyone who isn’t total newbie knows Nous Research And this framework kinda does fall within that space
165.
by zwaps 2y ago
I mean you have to have the climate and soil first
166.
by zwaps 2y ago
I really, really wish you would use a different citation format. Arbitrary numbers are really the least information. At least use last names and years, so I can have some idea which paper you are talking about without scrolling back and for
167.
by zwaps 2y ago
Many points in this blog post are just downright wrong. For example, the author doesn’t know what weight decay is. Otherwise this is a basic and incomplete Huggingface tutorial. I am sorry but not a good showing
168.
by zwaps 2y ago
SentencePiece is not a different algorithm to WordPiece or BPE, despite its naming. One of the main pulls of the SentencePiece library was the pre-tokenization being less reliant on white space and therefore more adaptable to non Western la
169.
by zwaps 2y ago
the improvement of gpt4 over even 3.5 is significant and for practical applications these improvements exceed what is indicated by benchmarks There are many things which earlier models maybe somewhat did, but only large models do reliably t
170.
by zwaps 2y ago
I think they mean that a random variable maps to a real number space. If it maps elsewhere, mathematicians like to call it a random element instead. https://en.m.wikipedia.org/wiki/Random_element
171.
by zwaps 2y ago
Is anyone else weirded out that they do not acknowledge Anthropic at all?
172.
by zwaps 2y ago
Awfully many coming soons in that article
173.
by zwaps 2y ago
Not clear value prop. Text "What makes Plan B different" reads: - Has polls - Has 360 surveys - Something about integrating tools Pretty sure other tools have all that
174.
by zwaps 2y ago
<< Static websites are fine, but dynamic ones are better.>> Nah
175.
by zwaps 2y ago
PEFT is not for adding knowledge, that's obvious
176.
by zwaps 2y ago
Which open source models are better than Claude 3?
177.
by zwaps 2y ago
In no way would I claim that the person does not have severe autism - but it is someone who consented to marriage. What I mean is a person where this would never ever happen.
178.
by zwaps 2y ago
Well yes, but also not really - I mean by no means are they as libertarian as the US eqivalent.
179.
by zwaps 2y ago
Germany hasn't had any mandatory service since more than a decade
180.
by zwaps 2y ago
Why not ask autistic people Have you ever even seen someone with severe autism?

More ›