Training on the rephrased test set is all you need: 13B models can reach GPT-4 performance in benchmarks with no contamination detectable by traditional methods

Covid-Plannedemic_ · 10 months ago

Training on the rephrased test set is all you need: 13B models can reach GPT-4 performance in benchmarks with no contamination detectable by traditional methods

its_just_andy · 10 months ago

if you’re interested in running your own models for any reason, you really should build your own evaluation dataset for the scenarios you care about.

at this point, all the public benchmarks are such a mess. Do you really care if the model you select has the highest MMLU? Or, do you care only that it’s the best-performing model for the scenarios you actually need?

Training on the rephrased test set is all you need: 13B models can reach GPT-4 performance in benchmarks with no contamination detectable by traditional methods

Training on the rephrased test set is all you need: 13B models can reach GPT-4 performance in benchmarks with no contamination detectable by traditional methods

Catch me if you can! How to beat GPT-4 with a 13B model | LMSYS Org