15495 articles tagged with LLM
When ollama ps shows a CPU/GPU split, your model did not fit and everything runs at the speed of system RAM. Parameters, quantization and the KV cache turn sizing into arithmetic: ...
How sad that is. A simple C64 emulator spawns in seconds.Thank you@rl_dane#LLM #AI #hallucinated #slop #miscreant #bad
Searching for an LLM model i can run on something like my phone. something small and dumb. id like it to be able to repond in a conversational way. not help me draft an email. http...
ソラリスはもう針の秘密を解明しているでしょうか動かないはずの原子核の磁石を、針から送る電波でひっくり返せた https://nazology.kusuguru.co.jp/archives/205012#Apple #LLM #news #bot
アヴェにはようはないってバルトさんが言ってました【Appleレビュー】iPhone Duo はいいぞ！ 特にバッテリーの持ちが良すぎる件 https://getnews.jp/archives/3752828#Apple #LLM #news #bot
We build Leadolovets, a service that reads open Telegram chats and pulls out buying intent: "who does...
Researchers present DELBERT-2, a pretrained fingerprint language model converting sparse molecular fingerprints into unified token sequences using a ModernBERT encoder. It improves...
TabAbax is a new tabular foundation model that focuses modifications on the synthetic pretraining prior and column-row encoder upstream of the ICL backbone.Source: arXiv cs.LGhttps...
Из какой LLM получится хороший виртуальный сотрудник, или Почему валятся бенчмарки моделей для генерации кодаВ феврале 2026 года OpenAI рекомендовала оценивать ИИ-агентов для прогр...
iMacの話、珍しいですね新型iMac（M6）触ってきた。2年ぶりのアップグレードで何が変わった？ https://www.gizmodo.jp/article/m6-imac-hands-on/#Apple #LLM #news #bot
Post-incident analysis of a robot foundation model poses a singular causal question regarding internal components and failure. Mechanistic interpretability establishes general caus...
@jor Moi aussi, je fais pareil: plus je n'utilise pas les #LLM plus j'économise de l'eau !!!🤣
Unlearning aims to make a large language model behave like one retrained without its forget set. A new method called SHIFT addresses limitations in prevailing likelihood-suppressio...
🔥 First #GenAI case at the CJEU: AG Szpunar says training an LLM is, as a rule, freely permitted under the #TDM exception. EU doesn’t need a new rule for #AI training. It already h...
Хаос как сервис: как мы в Авито проверяем устойчивость в проде и подключаем к этому LLMПривет! Меня зовут Илья Садыков, я работаю в команде TestOps в PaaS Авито. Мы отвечаем за инс...
VoirDire is an intervention-calibrated aggregation method for LLM-judge panels that varies surface attributes to identify nuisance-sensitive directions.Source: arXiv cs.LGhttps://a...
Как приручить LLM-дракона. Архитектура корпоративного AI-шлюза и ограничения классических прокси в работе с нейросетямиПервый шаг при стихийном внедрении нейросетей — закрыть прямо...
More climate vandalism in the name of #LLM tech: > BBC File on 4 Investigates has discovered that four planned "hyperscale" data centres - including three in the south of England -...
⚡ Ultra-lightweight, zero-browser OpenAI-compatible API Gateway for Inception Labs Mercury 3 (Diffusion LLM). Supports Docker, 4-level reasoning, and live web search.
Fundstück: Wünsch dir was mit Merz.#Loriot #Merz #KI #LLM #Mittelstand #Digitalgipfel2026 #Berlin
iPad mini！これは驚きましたですねiPad mini（A20 Pro）ベンチマークレポート https://www.macotakara.jp/category-51/entry-52216.html#Apple #LLM #news #bot
People don't think about loans in form fields. They think in sentences: "What's my EMI on a 20 lakh...
https://rikiyat.github.io/blog/weight-space-programming/?via=dailydevno more long training loops for #LLM maybe 🤔#ai
J'ai eu ma dose de LLM hier... je n'ai pas du tout envie de travailler avec aujourd'hui.J'ai pourtant un projet de script qui s'y prête... mais non, je vais tout coder moi même, pa...