I touched on the idea of sleeper agent LLMs at the end of my recent video, as a likely major security challenge for LLMs (perhaps more devious than prompt injection). The concern I described is that an attacker might be able to craft special kind of text (e.g. with a trigger
LLMS
-
Optimizing GPT Instructions for Enhanced Capabilities
By
–
This also made my GPT instructions 3 times smaller, meaning it can handle much more use cases! Try it here https://
chat.openai.com/g/g-Ft2u9p9PB-
testingcatalog-ai-news-features
… What's next?
– Personalisation? – Bigger app coverage? – New data sources? -
TestingCatalog GPT updated with custom actions for faster performance
By
–
TestingCatalog GPT just got updated! Now it relies on custom actions only – 10 times faster response!
– Realtime data access, no delays between publishing on TestingCatalog and availability on GPT.
– Added "Better AI features to test today" conversation starter. -
GPT-4 Turbo Vision: Higher Limits Enable New Creative Possibilities
By
–
I’m really excited about what GPT-4 Turbo with Vision enables, and can’t wait to see what you can build with these higher limits!
-
NeurIPS Podcast: Large-Scale Model Research with Joel Hestness
By
–
The latest @latentspacepod podcast interviewed our principal scientist Joel Hestness at NeurIPS covering all the large scale model work we are doing. Check it out!
-
Performance impact quantization 16-bit versus 4-bit models
By
–
what's the performance dropoff from 16 bits to 4?
-
AI Chatbots Hallucinate Legal Answers, Risks for Low-Income Users
By
–
Popular AI chatbots frequently hallucinate when answering legal questions, new research from Stanford RegLab and HAI finds. This poses special risks for people using the technology because they can’t afford a human lawyer. (via @blaw
) -
LLMs Excel at Code Generation and Statistical Analysis Tasks
By
–
hahah this one is kind of a stretch but i've tried telling it what i'm working on and asking for advice. usually it generates jargon in the right area that doesn't make any sense. but it's really good for making figures, running t-tests, writing evaluation code, etc.
-
DSPy: Open-Source Framework for Large Language Model Applications
By
–
DSPy: https://
github.com/stanfordnlp/ds
py
… -
Build hundreds of GPTs instantly with powerful templates toolkit
By
–
If you want to build hundreds of GPTs in one click using powerful templates, consider visiting my GPT Builder Toolkit: – 100+ Mega-Instructions for GPT Building
– How-to Guides & Tips
– Lifetime Access & Updates Get it before it’s gone: https://
godofprompt.ai/custom-gpt-too
lkit
…