jdiff

2,354 karmajoined hace 5 años

Educator, creative coder, slowly burning out.

@different55 at github.com

comments

anteayerdiscuss

I explicitly prompt my small local models to channel that kind of energy. It's one of few ways to get them to just spit it out without yapping on about temperature- and humidity-appropriate activities when I just asked if it'll rain this week.

You do have to do it carefully and not literally say "Star Trek" or else it'll start yapping about main shields being at full strength.

hace 4 díasdiscuss

Even in real pregnancies, it's estimated that as many as three quarters of all fertilized eggs fail to implant in the uterus. If one was of the opinion that life begins at conception, one should certainly be aware of that silent holocaust that has and always will eclipse all others for all of human existence.

hace 5 díasdiscuss

No, I'm satisfied with my own experience. If you would like to see data, take action towards those ends.

hace 5 díasdiscuss

Oh I certainly believe them. And on some level I admire the dogfooding. It just seems awful foolish of them.

hace 6 díasdiscuss

The minimal human supervision for prompts seems like a pretty silly take. LLMs are still pretty bad at creating good prompts for LLMs. Give it a benchmark and some feedback and it can brute force it, but far less effectively. And I thought the point of using LLMs for development was increased efficiency.

hace 6 díasdiscuss

Honesty has become a big tell, especially if its brutal. I'm not sure if it's getting worse or if I'm becoming more sensitive, but if I'm being brutally honest I can barely stand using LLMs anymore with their tone and their cliches. It's a horrible nightmare fusion of socmed influencer and LinkedIn hustler that sounds freakish even if it were human, and the density of tropes that are accumulating feels ridiculous.

Didn't the foundries take action against those in the past? I don't see "delve" nearly as often anymore. Why are the models spiraling like this now?

What really grinds my gears is the constant need to guess what I'm doing and offer a million random follow-ups. I asked what the weather was like, I don't appreciate the 5 paragraphs of tokens burned on weather-appropriate activity suggestions.

hace 6 díasdiscuss

The obnoxious cliche is mine, although I wouldn't call it "regular speech" since I tacked on that dense blob of LLMisms intentionally.

hace 6 díasdiscuss

But in that case you have nobody but yourself to blame, and you can stabilize things yourself at any time by refraining from making any changes. You won't be surprised by a provider. Honestly? That's not just valuable—it's essential.

hace 6 díasdiscuss

An n-deep chain of monitors doesn't really have any defense that an (n-1)-deep chain of monitors has. None of them have the capacity to separate data and instructions. All you're doing is (in some ways) giving the model more rolls of the dice to catch what's going on, but the kind of dice and the needed values to roll are in the attacker's hands as much as yours.

hace 6 díasdiscuss

I imagine the challenge comes from recognizing that your model is trying to call a tool before it actually has and only constraining output then. Running a separate pass for an optionally-empty list of tools afterwards may work, but maybe constraining its output like that causes many spurious tool calls.

hace 7 díasdiscuss

I think they know it'll hurt goodwill. I think they don't think they have to care what other people think about them. I think they think they're too big to fail.

hace 7 díasdiscuss

Feather has a charging circuit on board, and its usb port has a channel carved out for it and is close to the edge

hace 7 díasdiscuss

Just a lithium battery.

hace 7 díasdiscuss

This screen is entirely independent, with its own power source, so unlike most bistable screens this one could also report when its connection with the Machine is lost, and in a different way than simply turning off.

hace 8 díasdiscuss

You're not modding a pre-built $1049 device. The faceplates are removable and swappable with no disassembly needed, and this fancy one connects via bluetooth and is powered via a battery. Entirely non-invasive.

hace 10 díasdiscuss

No, that's not how LLMs work. It's all probabilities, and that issue has only deepened with providers silently falling back to worse models if they suspect you might be distilling their models. If an LLM rolls a bad token that can tip the whole balance of the response into utter nonsense.

People use LLMs to do vulnerability scanning by throwing them repeatedly at a codebase. Depending on the run they return with nothing, with a false positive, with a true vulnerability. These are very different destinations when faced with the same problem, sometimes.

Since GPT2, people have been throwing a ton of crap at the wall just to pick out one nugget that's uncharacteristically more solid than the others. Honestly? It's not just possible—it's core to how they operate. And it always has been.

hace 10 díasdiscuss

Is this a joke you're attempting? You're raging out about someone else being so narcissistic for not letting you continue on with your whole entire life because they took a vacation?

hace 10 díasdiscuss

I don't even know how we got here. This isn't that deeply represented in the training data. Is this what RLHF hath wrought? A new dialect of English based on corporatespeak and influencers, two heavy-hitting bullshitters?

hace 11 díasdiscuss

Some of them, the more corporate or tech-focused ones like .ai or .inc or .tech or .llc. Very many of them are comparable within a dollar of .org.

hace 12 díasdiscuss

Facebook has their own sins crawling on their backs, yes.