The truth is out there, hidden in the data. Below are the latest test results from my home‑brew AI experiments.
| Date | Model | Resource used | Tests | Result | Notes |
|---|---|---|---|---|---|
| 2023-10‑05 | OpenAI GPT‑3.5 | 4×CPU (Intel i5 3.2 GHz) | Sentiment analysis on 1 k Reddit comments | View result | Good performance, minor bias |
| 2023-09‑20 | Google PaLM 2 | GPU (NVIDIA GTX 1080) | Code generation for a Python sorting algorithm | View result | Generated correct code, occasional inefficiency |
| 2023‑08‑15 | Meta LLaMA 2 7B | 8×CPU (AMD Ryzen 9 3.8 GHz) | Language modeling on Wikipedia snippets | View result | High perplexity, slower than expected |