Tootfinder

Opt-in global Mastodon full text search. Join the index!

@freelancer0111@norden.social
2024-11-27 00:08:07

Der LfD Niedersachsen hält die Diskussion um den Personenbezug in LLMs für sinnvoll, vertritt aber die Auffassung, dass ein mit pbD trainiertes System Daten mit Personenbezug enthält und ausgibt. Damit schließt er sich der Argumentation seiner Kollegin in SH an. Seine Konsequenz: Es braucht ein ermöglichendes KI-Datenschutz-Gesetz abweichend zur #DSGVO. My 2 Cent: Das greift in den Konsequenzen zu …

@txiotxio@mastodon.eus
2025-01-19 10:26:33

>If this is the way to superintelligence, it remai lemmy.dbzer0.com/post/32982890

The GPT Era Is Already Ending - Divisions by zero
>If this is the way to superintelligence, it remains a bizarre one. “This is back to a million monkeys typing for a million years generating the works of Shakespeare,” Emily Bender told me. But OpenAI’s technology effectively crunches those years down to seconds. A company blog [https://openai.com/index/learning-to-reason-with-llms/] boasts that an o1 model scored better than most humans on a recent coding test that allowed participants to submit 50 possible solutions to each problem—…

@theawely@mamot.fr
2024-12-13 18:48:37

Excited about the new xLSTM model release. There are many well-though designs compared to transformers: recurrence (which should allows composability), gating (like Mamba & LSTM which is based on, which allows time complexity independent of the input size), state tracking (unlike Mamba & transformers). For now, these advantage aren’t apparent on benchmarks, but most training techniques are secrets, and the recent advances of LLMs evidenced that they matter a lot.