Viewing profile — TheGeminon

In the paper DeepSeek just says they have ~800k responses that they used for the cold start data on R1, and are very vague about how they got it: > To collect such data, we have ex…

Maybe they’ll try to build up traffic to your site from those domains and then push to sell them to you/extort by removing the redirects?

It is also unavailable from Canada.

If you read a bit further he excludes instances like that and listed films with only a single (likely intentional) title drop.

> Okay, so those are the problems. What’s the solution? > If you need to perform a case mapping on a string, you can use LCMap­String­Ex with LCMAP_LOWERCASE or LCMAP_UPPERCASE, po…

Devaluing the new currency by adding lesser metals will also devalue existing currency that is "pure" as you aren't able to trust the value of the currency anymore, so the value of…

There are ways to gauge the confidence of the LLM (token probabilities over the response, generating multiple outputs and checking consistency), but yeah that’s outside the LLM its…

Democracies are more than capable of passing laws that violate the rights of minorities.

So this would be a worthwhile case if a youthful physicist was bringing the complaint? As per the article, they are arguing that seniors are particularly affected by climate change…

With 37,000 Palestinians marked as suspected militants, it would mean they expected up to 555,000 - 740,000 civilian casualties.

This is also reported on by The Guardian https://www.theguardian.com/world/2024/apr/03/israel-gaza-ai...

According to Wiktionary in English that didn’t appear until the 14th century (although that is earlier than the article suggests) https://en.wiktionary.org/wiki/heart

The ideal way is to release a security advisory that some CVE patches are coming with no details, so that upgrades can be prepped coordinated for a known patch day.

You just need to focus niche and upmarket, OpenAI is e.g. never going to make that "clone your chats and have your LLM-self go on pre-dates" app that went around Twitter.

It doesn’t need to be all or nothing though, humans are social creatures. If rates are similar, having 5-10% of your tribe being more exploratory could be advantageous, even if the…

Outperform is dependent on the RAG approach (and this would be a RAG approach anyways, you can already do this with smaller context sizes). A simplistic one, probably, but dumping …

Where are you getting that the company is worth $50m? The article says it’s a “$100 million-a-year” business

From my (limited) reading it seems like this is a UN treaty to allow for better co-operation between states when prosecuting/investigating cybercrime. The main issues I see is that…

That confused be a bit at first too, but that ranking seems to be for how widely they are used (not sure the exact metric), not the ranking of how good they are (which is the dbms …

There is a more detailed explanation at https://unsloth.ai/introducing

You don’t need to apply the license until after the public release, before then it would just be sharing code with your customers under whatever agreement you want

That kind of decoding is a bit different though. For one, the tokenization process makes encodings difficult to handle (unless it’s trained on a lot of pairs). This would be more a…

The problem is that it’s doing a search of your relatively niche problem, and probably getting pretty poor results. The text from the search is then more highly weighted than the b…

This seems like a great way to pull out session identity from cookies and sidestep a lot of the baggage there. I would caution framing it as a secure replacement for cookies, it’s …

They could also check the physical meter for that, or if they are wardriving for the signal they could likely pick up the presence/lack of WiFi/5G/LTE signals which would probably …