Comparing Top AI Models on the Double Slit Experiment: Insights from ChatGPT, GPT-2, Claude 3, and More

Cristian Nedelcu
17 min readJun 7, 2024

I’m using top AI models according to lmsys.org, including ChatGPT, GPT-2, Gemini 1.5 Pro API, Claude 3 Opus, Yi Large Preview, LLaMA 3 70B Instruct, Reka Core, Qwen 2 72B Instruct, and Command R Plus, as well as others like Perplexity, PI AI, and Character AI. The prompt I’m using to compare their results is: “What is the double slit experiment?

https://chat.lmsys.org/

ChatGPT 4o

The double-slit experiment is a famous and fundamental experiment in quantum mechanics that demonstrates the wave-particle duality of light and matter. Here’s a detailed explanation:

### Experimental Setup
1. **Light Source:** A coherent light source, such as a laser, emits light toward an opaque barrier with two parallel slits.
2. **Barrier with Slits:** The barrier has two closely spaced slits, which can be opened or closed individually.
3. **Screen:** A detection screen is placed behind the barrier to observe the pattern formed by the light that passes through the slits.

### Classical Prediction
In classical physics, one would expect that if light were purely a particle, it would travel in straight lines through the slits and produce two distinct spots on the…

--

--