How to get the most effective results from the new ChatGPT-o1 (2024/09/20)
How to get the most effective results from the new ChatGPT-o1 (2024/09/20)
Hosted by Dr. Alvaro Cintas
Your current plan does not include access to this workshop.
Timestamps
[00:03] Start with intro, housekeeping, and recording info
[02:00] Share why today’s topic matters and agenda
[03:45] Explain ChatGPT-o1 and GPT-5 name skip
[04:46] Break down Chain-of-Thought prompting and paper
[07:20] Compare standard vs CoT prompting with diagram
[10:00] Show how o1 “thinks before speaking” internally
[12:05] Introduce Mini, Preview, and Full o1 models
[17:35] Dive into benchmarks: math, coding, human eval
[21:00] Cover pricing, access, and API rate limits
[23:10] Run CoT demo: counting letters in “raspberry”
[24:55] Share prompt-writing tips and best practices
[29:35] Recommend XML tags for complex prompt clarity
[32:00]Link shared doc with prompts...
Your current plan does not include access to the rest of this workshop.
Comments (0)
You need an active subscription to comment.