#LLM | AI Hub

🧠 OpenAI descrive il sistema pubblicitario di ChatGPT: un LLM segue l'evoluzione dell'intento e coordina la selezione degli annunci. Anche non mostrarne nessuno può essere un esito...

OpenAI LLM

9

Mastodon discussion Oct 23

iPod来ましたですHappy 25th Anniversary, iPod https://www.macrumors.com/2026/10/23/ipod-25th-anniversary/#Apple #LLM #news #bot

LLM

30

Mastodon discussion Oct 23

🚀 I Built the Smallest LLM I Could to Understand How They WorkI use LLMs every day without knowing how they work, so I b...

🚀 I Built the Smallest LLM I Could to Understand How They WorkI use LLMs every day without knowing how they work, so I built the smallest one I could: a tiny transformer that write...

LLM

9

Mastodon discussion Oct 23

When an LLM agent's move fails, writing the move and the failure message into the transcript makes small models more lik...

When an LLM agent's move fails, writing the move and the failure message into the transcript makes small models more likely to repeat the failed action in tool calling.Source: arXi...

LLM

9

Mastodon discussion Oct 23

An LLM agent can retrieve a memory correctly yet use it incorrectly after its validity state changes. We study governanc...

An LLM agent can retrieve a memory correctly yet use it incorrectly after its validity state changes. We study governance-state corruption faults that alter which retained records ...

LLM

9

Mastodon discussion Oct 23

Почему мы забрали у LLM браузер: 8 выводов после года работы над расширением для промокодовИзначально отдали GPT полный ...

Почему мы забрали у LLM браузер: 8 выводов после года работы над расширением для промокодовИзначально отдали GPT полный контроль: вот тебе страница магазина, вот список промокодов,...

LLM

9

Mastodon discussion Oct 23

RT @LinuxNews_de: Linux 7.3 bricht Rekorde – auch dank KI linuxnews.de/linux-7-3-bric… #kernel #ki #ai #llm #linux #linu...

RT @LinuxNews_de: Linux 7.3 bricht Rekorde – auch dank KI linuxnews.de/linux-7-3-bric… #kernel #ki #ai #llm #linux #linuxnews t.co/cIX8FT3Sjy mehr auf Arint.info #Innovation #KI #L...

LLM

9

Mastodon discussion Oct 23

新、一筋縄ではいきませんよゲームの自動デバッグも可能に、マウスコンピューターが示すローカルAIの活用法 https://news.mynavi.jp/techplus/article/20261023-5178901/#Apple #LLM...

新、一筋縄ではいきませんよゲームの自動デバッグも可能に、マウスコンピューターが示すローカルAIの活用法 https://news.mynavi.jp/techplus/article/20261023-5178901/#Apple #LLM #news #bot

LLM

27

Mastodon discussion Oct 23

Сколько на самом деле стоит запрос к LLM: формулы, кэш промптов, reasoning-токены и длинный контекстОбычно стоимость зап...

Сколько на самом деле стоит запрос к LLM: формулы, кэш промптов, reasoning-токены и длинный контекстОбычно стоимость запроса к языковой модели прикидывают так: открывают прайс, вид...

LLM

9

Dev.to tutorial Oct 23

Making LLM-as-Judge Scoring Reproducible and Trustworthy

A practical guide to fixed rubrics, structured output, temperature 0, and a deterministic mock judge so your LLM-based evals stay stable enough to gate CI.

LLM

12

Mastodon discussion Oct 23

🚀 Fastest-growing AI projects today1. The growth driven not only by the increasing complexity of LLM-based applications ...

🚀 Fastest-growing AI projects today1. The growth driven not only by the increasing complexity of LLM-based applications but a...2. The Nokia Applied Research team’s AnyJev reposito...

LLM

9

Mastodon discussion Oct 23

Cost per request is a misleading number for LLM apps.A cheap call that fails a quality check and gets retried three time...

Cost per request is a misleading number for LLM apps.A cheap call that fails a quality check and gets retried three times can cost more than one premium call that passes. Count cos...

LLM

9

Mastodon discussion Oct 23

Researchers introduced the SC-matrix to distinguish whether a multilingual LLM is sycophantic or lacks competence when a...

Researchers introduced the SC-matrix to distinguish whether a multilingual LLM is sycophantic or lacks competence when agreeing with a false premise. Results show multilingual syco...

LLM

9

Mastodon discussion Oct 23

LLM-беседы: Интеграция данных по оси «боли»Среда ( medium ) и-слоя в теории LLM-бесед как лумановских систем нового типа...

LLM-беседы: Интеграция данных по оси «боли»Среда ( medium ) и-слоя в теории LLM-бесед как лумановских систем нового типа ( Krupkin, 2026b ) не исчерпывается когнитивными, дискурсив...

LLM

9

Mastodon discussion Oct 23

IQuest-Q1 is a sparse Mixture-of-Experts foundation model with 320 billion parameters, of which 15 billion are activated...

IQuest-Q1 is a sparse Mixture-of-Experts foundation model with 320 billion parameters, of which 15 billion are activated.Source: arXiv cs.CLhttps://arxiv.org/abs/2610.30949#Machine...

LLM

9

Mastodon discussion Oct 23

Scientific claims from LLM social simulations should rely on robustness audits. Minor perturbations can cascade into mac...

Scientific claims from LLM social simulations should rely on robustness audits. Minor perturbations can cascade into macro-level outcomes, reflecting implementation artifacts.Sourc...

LLM

9

Mastodon discussion Oct 23