– Yes, across the board more compute = more accuracy (although without any confidence thresholds, you'll still get wild guesses or wrong answers)
LLMS
-
Compute Power Increases Model Confidence and Overconfidence Risks
By
–
– And good news! More compute = more confidence as well. (However, more compute made one model become more confident in wrong answers too. Something to keep an eye to make sure model confidence doesn’t tip into overconfidence.)
-
Teaching Models Confidence and Uncertainty Recognition
By
–
We know more compute results in higher accuracy, but are the models more confident those answers ARE accurate too? And how do we teach them when to say “I don’t know”? That’s what the research team wanted to find out.
-
Compute Budget and Confidence Thresholds Impact Model Math Performance
By
–
In the study, they measured how different combinations of compute budget and confidence thresholds (being at least 50% sure of the answer, etc.) affected models’ performance on a benchmark math test.
-
Johns Hopkins: LLMs Must Know When to Abstain
By
–
You can't just be right, you have to know you're right. Good advice for LLMs, according to new Johns Hopkins research. Sometimes no answer is better than a wrong one – life or death choices in medicine, for example, or big financial decisions.
-
Google Gemini copies ChatGPT, test reveals surprising results
By
–
The new #GoogleGemini is here and it's a TOTAL COPY of #ChatGPT. But what if the copy turned out to be better than the original? Here's my full test → https://
youtu.be/EY-cL9NTB8o Clearly… I wasn't expecting what I saw -
Discussing MCP Integrations for NotebookLM Users
By
–
Are you planing to integrate MCPs somehow? How realistic is that and which integrations do you think will be valuable for NotebookLM users?
-
Inquiry on DeepSearch and WebSearch Release ETA and Comparison to Gemini
By
–
What is the ETA for the DeepSearch and WebSearch release? Will it be similar to Deep Research on Gemini?
-
Long-Term Agentic Memory Course with LangGraph
By
–
New short course: Long-Term Agentic Memory with LangGraph. Learn to build an agent with long-term memory in this course developed in collaboration with @LangChainAI taught by its Co-Founder and CEO, @hwchase17!
— Andrew Ng (@AndrewYNg) 19 mars 2025
Personal assistance and productivity tasks have become important… pic.twitter.com/nTOCvoKUiiNew short course: Long-Term Agentic Memory with LangGraph. Learn to build an agent with long-term memory in this course developed in collaboration with @langchain taught by its Co-Founder and CEO, @hwchase17
! Personal assistance and productivity tasks have become important -
QwQ 32B Reasoning Model Now Live on SambaNova Cloud
By
–
Say Hello to QwQ 32B @Alibaba_Qwen
's QwQ 32B is NOW out of preview & LIVE on SambaNova Cloud. Running over 350 tps, this reasoning model approaches the capabilities of @deepseek_ai
's R1 671B in accuracy but with a much smaller model. Try it on SambaNova Cloud Today!