PVI Calculator

Starting the model… the first request after an idle period takes ~20–30s while the model loads onto the GPU.

Product Title

Product Description

Examples

Methodology. PVI(x) = H(Y | ∅) − H(Y | x), where H is V-entropy in bits from LoRA-adapted Llama-3.1-8B classifiers fine-tuned on Etsy product listings. The full model conditions on listing text; the null model sees only a blank input, providing the baseline entropy. Higher PVI means the listing reduces more uncertainty for an AI agent deciding whether to SELECT or PASS. A PVI near zero means the text adds little beyond the baseline; negative PVI means it is less informative than having no input at all.

Implementation. The base model (Llama-3.1-8B-Instruct) is quantized to 4-bit (NF4) and served with two LoRA adapters (weights) on a single T4 GPU via Modal. Anonymized request logs (IP address, inputs, and results) are retained for research purposes. Repeated queries are served from cache.