English
English Indonesia Türkçe Deutsch Español Português Pусский Français Italiano Nederlands Polski العربية ไทย Tiếng Việt 繁體中文 简体中文 日本語 한국어
curlyhacks1's profile picture. Exploring LLM knowledge manipulation and weird ideas. Founder of @neodynamics_ai (AI for Cyber defense). 👾 hacker of all the things tech

Andrea Villa

@curlyhacks1

Exploring LLM knowledge manipulation and weird ideas. Founder of @neodynamics_ai (AI for Cyber defense). 👾 hacker of all the things tech

243Posts 43Followers 89Following
Pinned
curlyhacks1's profile picture.
Andrea Villa
@curlyhacks1
Oct 28

TUTORIAL: how to obtain instruction-following (NOT base model) specialized model for your knowledge base / domain expertise / specific task [DATASET VERSION]. 🧵

1
0
2
0
407
curlyhacks1's profile picture.
Andrea Villa
@curlyhacks1
Nov 3

There's still no epistemic model of why a result is valid, but it is at least a very good "digger" of new mappings between ideas.

godofprompt's profile picture. 🔑 Sharing AI Prompts, Tips & Tricks. The Biggest Collection of AI Prompts & Guides for ChatGPT, Grok, Claude & Midjourney AI → https://t.co/vwZZ2VSfsN
God of Prompt
@godofprompt
Nov 3

Holy shit... researchers just built an AI scientist that can do actual end-to-end research 😳 It’s called 'Denario' a multi-agent system that literally thinks, plans, codes, experiments, and writes papers… all by itself. You give it a dataset or a vague question, and it will:…

godofprompt's tweet image. Holy shit... researchers just built an AI scientist that can do actual end-to-end research 😳 It’s called 'Denario' a multi-agent system that literally thinks, plans, codes, experiments, and writes papers… all by itself. You give it a dataset or a vague question, and it will:…
49
158
756
849
71K
0
0
0
1
108
curlyhacks1's profile picture.
Andrea Villa
@curlyhacks1
Oct 30

Great tutorial to reproduce R1 using GRPO. Outdated a little by now, but worth understanding / implementing @_philschmid philschmid.de/mini-deepseek-…

curlyhacks1's tweet card. Reproduce Deepseek R1 „aha moment“ and train an open model using reinforcement learning trying to teach it self-verification and search abilities all on its own to solve the Countdown Game.
Mini-R1: Reproduce Deepseek R1 „aha moment“ a RL tutorial
Source: philschmid.de
0
0
1
2
109
Andrea Villa reposted
QuanticASI's profile picture.
φ
@QuanticASI
Oct 29

Roger Penrose calculated that the probability of a universe like ours arising by chance is on the order of 1 in 10¹⁰¹²³, an inconceivably small number but you're here

QuanticASI's tweet image. Roger Penrose calculated that the probability of a universe like ours arising by chance is on the order of 1 in 10¹⁰¹²³, an inconceivably small number but you're here
410
439
2K
715
170K
curlyhacks1's profile picture.
Andrea Villa
@curlyhacks1
Oct 29

Dear algo: show this to entrepeneurs in AI and cyber space. Respond with "hi" to connect.

0
0
1
0
61
Andrea Villa reposted
jaseweston's profile picture.
Jason Weston
@jaseweston
Oct 29

🌶️SPICE: Self-Play in Corpus Environments🌶️ 📝: arxiv.org/abs/2510.24684 - Challenger creates tasks based on *corpora* - Reasoner solves them - Both trained together ⚔️ -> automatic curriculum! 🔥 Outperforms standard (ungrounded) self-play Grounding fixes hallucination & lack of…

jaseweston's tweet image. 🌶️SPICE: Self-Play in Corpus Environments🌶️ 📝: arxiv.org/abs/2510.24684 - Challenger creates tasks based on *corpora* - Reasoner solves them - Both trained together ⚔️ -> automatic curriculum! 🔥 Outperforms standard (ungrounded) self-play Grounding fixes hallucination & lack of…
7
55
319
243
75K

United States Trends

Something went wrong.


Something went wrong.