cfcf14

7 ms·

1.
by cfcf14 5mo ago
I would assume so too, so the costs would not be so substantial to Anthropic.
2.
by cfcf14 5mo ago
I'm curious as to why 4.7 seems obsessed with avoiding any actions that could help the user create or enhance malware. The system prompts seem similar on the matter, so I wonder if this is an early attempt by Anthropic to use steering
3.
by cfcf14 9mo ago
This makes me think it would be nice to see some kinda child of modern transformer architecture and neural ODEs. There was such interesting work a few years ago on how neural ode/pdes could be seen as a sort of continuous limit of laye
4.
by cfcf14 2y ago
The obvious next step here is to see how well this generalises to arbitrary inputs :)
5.
by cfcf14 2y ago
Did your read the paper? Do you have specific criticisms of their problem statement, methodology, or results? There is a growing body of research indicating that in fact, there _is_ a taxonomy of 'hallucinations', that they might
6.
by cfcf14 2y ago
AI detectors do not work. I have spoken with many people who think that the particular writing style of commercial LLMs (ChatGPT, Gemini, Claude) is the result of some intrinsic characteristic of LLMs - either the data or the architecture.
7.
by cfcf14 2y ago
So uh, things are not looking so good for actual physics these days, I gather?
8.
by cfcf14 2y ago
L-theanine (200mg) with around 100-150mg of caffeine has an extremely noticeable, positive effect on my ability to focus, feeling of "well-situatedness", and overall calmness. L-theanine by itself doesn't seem to do much. Caf
9.
by cfcf14 2y ago
It's not - the FTC released a statement on this very topic a few months ago: https://www.ftc.gov/business-guidance/blog/2024/03/price-fix...
10.
by cfcf14 2y ago
Police in large American cities are not likely to be of much assistance in this situation. Assuming they attend at all, I would expect them to not understand the nature of the issue and probably proceed to make it much worse.
11.
by cfcf14 2y ago
After reading the paper, I'm really unsure what the novel contribution is. It feels like they're attempting to rebrand well-understood concepts within various fields (control systems theory, etc). The provided mathematical definit
12.
by cfcf14 3y ago
This is really funny - it's bordering on truly absurd, almost incomprehensible madness to consider doing this seriously. I can't think of a single property you'd desire in a control system (state observability, auditability,
13.
by cfcf14 3y ago
Was fun while it lasted! Will be interesting watching the internal story of the original lab unfold as it all becomes public eventually.
14.
by cfcf14 3y ago
Yeah - more or less. I still use it sometimes for trivial stuff like giving me recipe or travel inspiration (using the web search API), but I haven't been using it for any sort of algorithms/coding stuff. It's useful for task
15.
by cfcf14 3y ago
Some strange claims in this post. The reddit post datasets are already 'out there' in the wild, and I'm fairly certain every other major LLM release has used their data. Also - did Midjourney "steal" DALLE-2's
16.
by cfcf14 3y ago
Yeah, definitely. Combination of expert-system gating (some requests probably get routed to weaker models), distillation (for performance/cost), and RLHF lobotomization.
17.
by cfcf14 3y ago
It's turtles all the way down, except for the final turtle, which is Fortran...
18.
by cfcf14 3y ago
Lilian Weng's blog is my go-to example for an extremely high quality tech blog, it's truly remarkable how consistently excellent each post is. The only downside is the sadness I feel for being incapable of producing content even r
19.
by cfcf14 4y ago
This is 100% related to (suspected) fraud, anti money laundering, or other types of financial/political sanctions. You may be 100% innocent, but they will never disclose any information to you about their reasoning (and in fact it is i
20.
by cfcf14 4y ago
Absolutely not.
21.
by cfcf14 4y ago
You have to prompt it correctly, non-instruction-aligned models don't behave like agent simulators by default.
22.
by cfcf14 4y ago
Amazing post, agree with everything you've said. I've always felt that the problems with advanced MCMC methods (HMC, RM-MC, etc) are even more painful when one looks at approximate bayesian methods - ADVI (variational approximatio
23.
by cfcf14 4y ago
I wonder whether Bing has been tuned via RLHF to have this personality (over the boring one of ChatGPT); perhaps Microsoft felt it would drive engagement and hype. Alternately - maybe this is the result of less RLHF. Maybe all large model
24.
by cfcf14 4y ago
This was a really reasonable and interesting post by Stephen. I'm excited to see what the integration between an associative based model like GPT and a symbolic one like WA might bring.
25.
by cfcf14 4y ago
It being from Schmidhuber's lab makes it dramatically more credible, in my views. They've been practically a decade ahead of everybody for ages now from a theoretical point of view.
26.
by cfcf14 4y ago
He doesn't say this directly, but I suppose his comment here: "The whole 'do an all-nighter to get the paper in the day before the dealine' is something that should have gone out the window after highschool. Not for kern
27.
by cfcf14 4y ago
I discovered a while ago that you can ask GPT-3 for the output of even extremely obfuscated javascript code and it will produce the correct results most of the time.
28.
by cfcf14 4y ago
Couple reasons come to mind: 1) The institutional apparatus required to maintain the monarch as head of state is reasonably complex and expensive, both from a legal point of view and in a financial sense. The governor general is an unelecte
29.
by cfcf14 4y ago
As I Canadian I would strongly support removing the English monarchy as our head of state. I hope other countries do so as well.
30.
by cfcf14 4y ago
- Unreliable and buggy core libraries (basic fp math, stats, autodiff) which seems to stem from an academic-style disinterest in focusing on the boring bits of language foundations - 1 indexed arrays - extremely slow start up times (1 min??

More ›