Don’t you think we’ll just move the goal posts when we get there? Already the “Turing test” seems to be shrugged off as not important
RESEARCH
-
Is OpenAI’s o1 Model a Real Breakthrough or Just GPT-4 Tuning?
By
–
There’s a mixed reaction to o1 — with critics saying it's less of a new model and more like GPT-4 with fine tuning. So, is this new o1 model by OpenAI a miraculous breakthrough, or not? 1. It’s pretty cool that chain of thought + search allows us to squeeze out even more
-
Transforming the World Through Belief in the Impossible
By
–
La fuerza transformadora de la creencia en lo imposible. Las personas que se atreven a soñar en grande, que desafían los límites de lo establecido y piensan de manera diferente, son las que logran cambiar el mundo. La innovación y el cambio comienzan con el coraje de creer que
-
Reviewer Concerns About Area Chairs at Major ML Conferences
By
–
Agree. When I was a reviewer for NeurICMLR, I felt like more often than not, ACs were out to lunch. I anecdotally hear the same from current reviewers.
-
Understanding OpenAI’s o1 Model and Latest AI Developments
By
–
What is o1 – OpenAI’s latest model… and how is it different?
— Amitav Bhattacharjee (@bamitav) 15 septembre 2024
pic.twitter.com/zSsxCwTHlA#GPT #XAI #LLM #LLMs #GenAI #GenerativeAI #OpenAI #ChatGPT #GPT #tech #AI #technology #artificalintelligence #Chatbot #MetaAI #llama31 #Llama3 #GROK2 #GROK2AI #GROK #chatgpt4 #Gemini…What is o1 – OpenAI’s latest model… and how is it different? #GPT #XAI #LLM #LLMs #GenAI #GenerativeAI #OpenAI #ChatGPT #GPT #tech #AI #technology #artificalintelligence #Chatbot #MetaAI #llama31 #Llama3 #GROK2 #GROK2AI #GROK #chatgpt4 #Gemini
-
LLM Peak Performance: Training, Inference, System Optimization
By
–
10). Achieving Peak Performance for LLMs – a systematic review of methods for improving and speeding up LLMs from three points of view: training, inference, and system serving; summarizes the latest optimization and acceleration strategies around training, hardware, scalability,
-
Flash-Sigmoid: Hardware-Efficient Attention 17% Faster
By
–
9). Theory, Analysis, and Best Practices for Sigmoid Self-Attention – proposes Flash-Sigmoid, a hardware-aware and memory-efficient implementation of sigmoid attention; it yields up to a 17% inference kernel speed-up over FlashAttention-2 on H100 GPUs; show that SigmoidAttn
-
Can LLMs Generate Novel Scientific Research Ideas
By
–
8). Can LLMs Unlock Novel Scientific Research Ideas – investigates whether LLM can generate novel scientific research ideas; reports that Claude and GPT models tend to align more with the author's perspectives on future research ideas; this is measured across different domains
-
Small Language Models Role Applications in LLM Era
By
–
6). The Role of Small Language Models in the LLM Era – closely examines the relationship between LLMs and SLMs; common applications of SLMs include data curation, training stronger models, efficient inference, evaluators, retrievers, and much more; includes insights for
-
LLaMa-Omni: Low-Latency Speech-to-Speech LLM Model Architecture
By
–
7). LLaMa-Omni – a model architecture for low-latency speech interaction with LLMs; it is based on Llama-3.1-8B-Instruct and can simultaneously generate both text and speech responses given speech instructions; responses can be generated with a response latency as low as 226ms…