In both the prompt and response above, each letter in “strawberry” is assigned by the GPT-4o tokenizer to a separate single-character token, instead of being combined into subword-length tokens like “st” “raw” “berry” etc.:
LLMS
-
Dispelling letter-counting myth by asking ChatGPT about strawberry
By
–
Dispelling a popular myth about letter-counting issues in LLMs by asking ChatGPT: h·o·w m·a·n·y R’s a·r·e i·n “s·t·r·a·w·b·e·r·r·y”?
-
Software Optimizations Maintain FP16 Weight Precision
By
–
just software optimizations. still FP16 weights.
-
Compiler Optimizations and Memory Management for FP16 Models
By
–
compiler optimizations, better memory management etc. same FP16 weights.
-
6 Amplifying AI Tools This Week With OpenAI o1
By
–
Missed our latest newsletter? No worries! We’ve got you covered with 6 Amplifying AI Tools for This Week Whether it's OpenAI’s groundbreaking o1 models or tools to supercharge your productivity, here’s a quick rundown of what you need to know. Let’s dive in! #AI #Tech
-
OpenAI o1 Models Outperform Experts; Microsoft Launches Copilot Agents
By
–
Big News: OpenAI’s new o1 models (aka ‘Project Strawberry’) outperform experts on PhD-level problems. Microsoft boosted 365 apps with Copilot Agents & launched open-source benchmarks for AI. Explore these & more! #AI #Innovation
-
Models Struggle with Instruction Following and Hallucinated Information
By
–
In our error analysis, we found that most models continued to struggle with: – Final Answer Missing Information
– Hallucinated Information This indicates a widespread issue with precise instruction following, especially in prompts requiring detailed information and intermediary -
SEAL Launches Agentic Tool Use Leaderboards, OpenAI o1 Ranks
By
–
SEAL Leaderboard Update Today, we're launching 2 new Agent leaderboards:
– Agentic Tool Use (Chat)
– Agentic Tool Use Enterprise) We are also adding OpenAI o1 to the Agentic Tool Use leaderboards, ranking at: #1 on Enterprise Tool Use
#3 on Chat Tool Use (read on) -
GRADIO-CODER: Instruction Fine-Tuned StarCoder for Gradio Development
By
–
GRADIO-CODER model: https://
huggingface.co/matrixglitch/G
RADIO-CODER
… This language model is the version 0.0 of a Gradio Coding Assistant. It is an instruction fine-tuned version of StarCoder that is designed to provide assistance to developers who use gradio. -
Unified Memory Management Across AI Models Implementation
By
–
I know, I need that so I can deal with unified memory across models