One of the best ways to improve LLM performance is to ask it to “think aloud” (there are various techniques for doing this, including Chain of Thought). This also helps establish clearly the AIs plans. This paper suggests that, in some cases, the AI can plan without revealing it
RESEARCH
-
China unveils Vidu, AI video generator rivaling OpenAI’s Sora
By
–
Overnight, China released their own version of OpenAI’s Sora:
— AI Breakfast (@AiBreakfast) 27 avril 2024
AI video generator “Vidu” can create 16 second clips at 1080p:pic.twitter.com/LkACnl1M2oOvernight, China released their own version of OpenAI’s Sora: AI video generator “Vidu” can create 16 second clips at 1080p:
-
Parameter-Efficient Fine-Tuning Methods for Large Language Models
By
–
Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey PEFT algorithms are useful for dealing with LLMs with high parameter counts, as even fine-tuning these models from scratch can be computationally expensive and resource-intensive. https://
arxiv.org/abs/2403.14608 -
Bitsphi Uses Deep Learning to Identify ADHD From Brain Signals
By
–
Startup Bitsphi uses #DeepLearning to process brain signals with an algorithm that can identify ADHD!
-
Editing Latent Space Podcast with Nous Research and WebSim
By
–
[Livestream] Editing the next @latentspacepod: @nousresearch, @websim_ai, @plinz https://t.co/5djN0Bmvrw
— swyx 🐣 (@swyx) 27 avril 2024[Livestream] Editing the next @latentspacepod
: @nousresearch
, @websim_ai
, @plinz -
AI Systems Show Impressive Reliability Compared to Humans
By
–
But when you think of them like humans, they are impressively reliable.
-
Advanced AI Systems Struggle with Basic Instruction Following
By
–
It's just so weird and unintuitive having super-advanced computers that can't reliably do the things that we tell them to do
-
Large Context Windows Boost LLM Performance Through Many-Shot Learning
By
–
Very large context windows may extend the capabilities of LLMs because you can give them hundreds of examples on how to solve a problem (many shot learning). This paper from Google finds significant performance boosts from many shot, even when the AI generates its own examples.
-
Compound AI Systems Outperform Single Models Approach
By
–
Compound AI systems bring models, feedback, external knowledge together; going to be much better than just single models/ single shot approaches!