Multiscale Policy Optimization for Long-Horizon LLM Agents
Long-horizon LLM agents interact with their environments through complete assistant turns while generating each action token by token. We observe that critic values vary primarily ...
#LLM
15512 articles tagged with LLM
OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?https://minimaxir.com/2026/10/cheap-llms...
OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?https://minimaxir.com/2026/10/cheap-llms/Comments: https://news.ycombinator.com/item?id=50236102#Hac...
Nanbeige LLM Lab has released Nanbeige4.3-33B, an open Apache 2.0 model aimed at agentic coding and professional cowork ...
Nanbeige LLM Lab has released Nanbeige4.3-33B, an open Apache 2.0 model aimed at agentic coding and professional cowork tasks. The 33B dense model carries 512K context, scores 80.9...
I've been wanting to learn how an #LLM works and built https://llm.kitchen/ to finally do that.
I've been wanting to learn how an #LLM works and built https://llm.kitchen/ to finally do that.
🔥 Hiring Alert! | Web Research Specialist (Llm)Turing, remote contract, 8 weeks, 40 hrs/week✅ Pay discussed during scree...
🔥 Hiring Alert! | Web Research Specialist (Llm)Turing, remote contract, 8 weeks, 40 hrs/week✅ Pay discussed during screening✅ Building benchmark questions AI agents cannot solve✅ T...
So guess what, looks like at least for now my new job isn't approving me for an #LLM license, I'm going to have to do al...
So guess what, looks like at least for now my new job isn't approving me for an #LLM license, I'm going to have to do all my programming by hand 🥳#AI
OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?
OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?
Amazon、バルトさんが好きそうです週末にスマートホームをオープンソース化--Googleから「Siri」へ乗り換えた話 https://japan.zdnet.com/article/35253722/#Apple #LLM #news...
Amazon、バルトさんが好きそうです週末にスマートホームをオープンソース化--Googleから「Siri」へ乗り換えた話 https://japan.zdnet.com/article/35253722/#Apple #LLM #news #bot
RE: https://mastodon.social/@fllics/117501990016777072#AIsafety #LLM #CNN #AIreasoning #AIAgents #ComputerScience #Techn...
RE: https://mastodon.social/@fllics/117501990016777072#AIsafety #LLM #CNN #AIreasoning #AIAgents #ComputerScience #Technology #NeuralNetworks #frontierAI
instead of clicking "i agree" on a terms of service update that takes up 134 KB i fed it into a local #llm and got a sum...
instead of clicking "i agree" on a terms of service update that takes up 134 KB i fed it into a local #llm and got a summary for my use casei feel pretty good about this
Day 08/90: Who Evaluates the Evaluator? When LLM Judges Get It Wrong
You update a prompt, generate a new answer, and ask another LLM to evaluate it. _ The score goes from...
Evals before prompts: a small eval harness for an LLM feature
A small eval harness for an LLM feature in TypeScript: golden cases, exact checks, an LLM judge with a pass/fail rubric, a CI gate and a tracked score.
Want to challenge trust in an open source project?Discuss your #LLM policy in secret.
Want to challenge trust in an open source project?Discuss your #LLM policy in secret.
Testing turing-unlock on my triple 2080ti LLM server. Will report back. #llm #homelab #localai
Testing turing-unlock on my triple 2080ti LLM server. Will report back. #llm #homelab #localai
The A in the current iteration of #AI really needs to stand for Augmented. Asking an #LLM to help us do something when w...
The A in the current iteration of #AI really needs to stand for Augmented. Asking an #LLM to help us do something when we don’t have the first clue what we’re asking for is not a g...
A 10-year-old CNN outperforms frontier LLMs for skin cancer diagnosishttps://lynchlab.net/llm-skin-cancerComments: https...
A 10-year-old CNN outperforms frontier LLMs for skin cancer diagnosishttps://lynchlab.net/llm-skin-cancerComments: https://news.ycombinator.com/item?id=50233779#HackerNews #CNN #Sk...
It seems that the latter has been amplified by the #genAI #LLM craze.Maybe there's another thing at play here as well: I...
It seems that the latter has been amplified by the #genAI #LLM craze.Maybe there's another thing at play here as well: I cannot imagine being in love with #SiliconValley (as a plac...
Just jump scared myself the #LLM used my full name and email address.Until I realized that in the task I pasted my full ...
Just jump scared myself the #LLM used my full name and email address.Until I realized that in the task I pasted my full configuration.nix and tasked it with putting it into separat...
A short write-up: Existentialist Slop.https://buttondown.com/hypernoetic/archive/existentialist-slop/#Blog #AI #LLM #The...
A short write-up: Existentialist Slop.https://buttondown.com/hypernoetic/archive/existentialist-slop/#Blog #AI #LLM #TheRegister #Camus #Derrida #Mistral
🧠 Researchers use a language model to simulate fake tool outputs when red-teaming AI agent monitoring systems. The appro...
🧠 Researchers use a language model to simulate fake tool outputs when red-teaming AI agent monitoring systems. The approach tests whether monitoring systems can detect when agents ...
Previously google had all our information due to continuous search of everything. Now #LLM companies have many of them. ...
Previously google had all our information due to continuous search of everything. Now #LLM companies have many of them. :-/I hope #localLLM progress would get faster so our normal ...
Chcete lokální LLM kvůli soukromí, ale chybí vám nástroje a práce se soubory? 💻🔐V ToolDurusu připojíte svůj model přímo ...
Chcete lokální LLM kvůli soukromí, ale chybí vám nástroje a práce se soubory? 💻🔐V ToolDurusu připojíte svůj model přímo do našeho chatu.🚀 Rychlé nastavení přes náš konektor🚀 Soukro...
How to Build an Eval Plan for an LLM Feature Before You Ship It
Most LLM features get approved the same way. Someone tries a dozen inputs, the outputs look...
Stanford and UC Berkeley researchers have released TokenGate, a drop-in module that teaches an existing LLM to prune uni...
Stanford and UC Berkeley researchers have released TokenGate, a drop-in module that teaches an existing LLM to prune unimportant tokens from long prompts. In tests up to 128k token...