Inducing Unprompted Misalignment in LLMs — LessWrong
Emergent Instrumental Reasoning Without Explicit Goals TL;DR: LLMs can act and scheme without being told to do so. This is bad. …
AI for Data Journalism: demonstrating what we can do with this stuff right now
I gave a talk last month at the Story Discovery at Scale data journalism conference hosted at Stanford by Big Local News. My brief was to go deep into the …
Limitless
Go beyond your mind’s limitations: Personalized AI powered by what you’ve seen, said, and heard.
Stuff we figured out about AI in 2023
2023 was the breakthrough year for Large Language Models (LLMs). I think it’s OK to call these AI—they’re the latest and (currently) most interesting development in the academic field of …
OpenAI Begins Tackling ChatGPT Data Leak Vulnerability · Embrace The Red
Good news. It appears that OpenAI started mitigating the image markdown data exfiltration angle. It remains vulnerable, but it's great to see a few first actions being taken to mitigate the problem.
Clap Or AI Gets It - Aftermath
Drama over YouTuber Marques Brownlee's review of the Humane AI pin shows how much AI is propped up by hype.
GitHub - SpellcraftAI/turing: Turing machines, Rule 110, and A::B reversal using Claude 3 Opus.
Turing machines, Rule 110, and A::B reversal using Claude 3 Opus. - SpellcraftAI/turing
BarGPT, AI Generated Cocktail Recipes
Create AI generated cocktail recipes based on your tastes, ingredients on hand or any other ideas you have.
Attention in transformers, visually explained | Chapter 6, Deep Learning
Demystifying attention, the key mechanism inside transformers and LLMs.Instead of sponsored ad reads, these lessons are funded directly by viewers: https://3...
Is Google's AI Actually Discovering 'Millions of New Materials?'
"In the DeepMind paper there are many examples of predicted materials that are clearly nonsensical."
Semgrep: AutoFixes using LLMs · Choly's Blog
A solid pattern to build LLM Applications (feat. Claude)
The thing about modern AI development - both developing things with AI and developing AI things - is that you often need to know the right magic incantations...
infiniflow/ragflow: RAGFlow is an open-source RAG (Retrieval-Augmented Generation) engine based on deep document understanding.
RAGFlow is an open-source RAG (Retrieval-Augmented Generation) engine based on deep document understanding. - infiniflow/ragflow
Building files-to-prompt entirely using Claude 3 Opus
files-to-prompt is a new tool I built to help me pipe several files at once into prompts to LLMs such as Claude and GPT-4. When combined with my LLM command-line …
AI in Drug Discovery - A Highly Opinionated Literature Review (Part III)
Following up on Part I and Part II, the third post in this series is a collection of review articles published in 2023 that I found helpfu...
Command R
Command R is a conversational model that excels in language tasks and supports multiple languages, making it ideal for coding use cases that require instruction models. It responds well to preambles that follow a specific structure and format, enhancing its performance.
nilsherzig/LLocalSearch: LLocalSearch is a completely locally running search aggregator using LLM Agents. The user can ask a question and the system will use a chain of LLMs to find the answer. The user can see the progress of the agents and the final answer. No OpenAI or Google API keys are needed.
LLocalSearch is a completely locally running search aggregator using LLM Agents. The user can ask a question and the system will use a chain of LLMs to find the answer. The user can see the progres...
Top Israeli spy chief exposes his true identity in online security lapse
Exclusive: Yossi Sariel unmasked as head of Unit 8200 and architect of AI strategy after book written under pen name reveals his Google account
‘Lavender’: The AI machine directing Israel’s bombing spree in Gaza
The Israeli army has marked tens of thousands of Gazans as suspects for assassination, using an AI targeting system with little human oversight and a permissive policy for casualties, +972 and Local Call reveal.
The cost of AI reasoning over time.
As time progresses, AI models are achieving higher reasoning accuracy while their associated costs continue to drastically decrease. What does it mean for our future?
Consensus: AI Search Engine for Research
ChatGPT for Research. Consensus is an AI-powered search engine that finds and summarizes scientific research papers. Just ask a question!
Add WebPilot to your GPTs
AI powered Search, access any online information, and generate very long content
BeeTrove - OpenAI GPTs Dataset
summary of AI agents for OpenAI
plandex-ai/plandex: An AI coding engine for complex tasks
An AI coding engine for complex tasks. Contribute to plandex-ai/plandex development by creating an account on GitHub.
But what is a GPT? Visual intro to Transformers | Deep learning, chapter 5
An introduction to transformers and their prerequisitesEarly view of the next chapter for patrons: https://3b1b.co/early-attentionOther recommended resources...
The Deodorant AI Spokesmodel Is a Real Person, Sort Of
An AI-generated ad for deodorant wipes has driven the internet insane.
Choose Your Weapon: Survival Strategies for Depressed AI Academics
infiniflow/ragflow: RAGFlow is an open-source RAG (Retrieval-Augmented Generation) engine based on deep document understanding.
RAGFlow is an open-source RAG (Retrieval-Augmented Generation) engine based on deep document understanding. - infiniflow/ragflow
CrewAI - Building a Custom Crew
In this video I go through an example of building a Custom Crew with CrewAI and compare the Sequential vs Hierarchical process methodsCODESequential Colab: h...
DSPy: Not Your Average Prompt Engineering
Heads up, Bay Area guys ditched their AVP already and buzz about DSPy now. Could DSPy be the fresh go-to framework for prompt engineering after LangChain and LlamaIndex?