I would love to see a dataset of English word-pairs rated by their severity (i.e., how much "very" applies to modify the first to get the second on a scale from

I would love to see a dataset of English word-pairs rated by their severity (i.e., how much "very" applies to modify the first to get the second on a scale from -1 to 1). For example, "quiet" would have "silent" rated as 1, "roomy" might have "spacious" listed close to 0, and "gorgeous" would have "pretty" as negative.