3 articles

📰 Dev.to · Shuntaro Okuma

Articles from Dev.to · Shuntaro Okuma · 3 articles · Updated every 3 hours · View all reads

All ⚡ AI Lessons (9392) ArXiv cs.AIDev.to · FORUM WEBForbes InnovationDev.to AIOpenAI NewsHugging Face Blog
I Tested 12 LLMs With Few-Shot Examples. The Results Were Not What I Expected.
Dev.to · Shuntaro Okuma 2w ago
I Tested 12 LLMs With Few-Shot Examples. The Results Were Not What I Expected.
In a previous article, I tested 8 models across 4 tasks and reported on "few-shot collapse" — cases...
Read article → + Playlist
How I Measure My Dify Chatbot Quality with Scenario Testing
Dev.to · Shuntaro Okuma 2w ago
How I Measure My Dify Chatbot Quality with Scenario Testing
What I did I designed multi-turn conversation scenarios for a Dify chatbot, ran them...
Read article → + Playlist
When More Examples Make Your LLM Worse: Discovering Few-Shot Collapse
Dev.to · Shuntaro Okuma 1mo ago
When More Examples Make Your LLM Worse: Discovering Few-Shot Collapse
I found that adding few-shot examples can actively degrade LLM performance. Here are three patterns I discovered and an open-source tool to detect them.
Read article → + Playlist