Description
Calculate second-minus-first log odds from two supplied sigmoid outputs.
A sigmoid output strictly between zero and one has log odds ln(p/(1-p)). Comparing log odds gives a signed contrast on a natural-log scale, with zero for equal outputs. Supplied values need their own model and calibration context before the contrast has an empirical interpretation.
Calculate second-minus-first log odds from two supplied sigmoid outputs.
When to use this tool
Use Sigmoid Output Log Odds Contrast when you already have firstOutput, secondOutput and need to calculate second-minus-first log odds from two supplied sigmoid outputs.
How to use the tool and how it works
- Enter firstOutput, secondOutput. Use the units and conventions below.
- Run the tool. Correct any format, range, or cross-field input error before interpreting the output.
- Read logOddsContrast. Use the output convention below.
Require each output strictly between zero and one; return [ln(second)-log1p(-second)]-[ln(first)-log1p(-first)].1
| Input field | Meaning and units | Accepted input |
|---|---|---|
| firstOutput | First Output | Required. Finite number; strictly above 0; strictly below 1. |
| secondOutput | Second Output | Required. Finite number; strictly above 0; strictly below 1. |
| Output field | Interpretation | Returned shape |
|---|---|---|
| logOddsContrast | Natural log odds(secondOutput)-natural log odds(firstOutput), not first-minus-second or a probability difference. | Required. Finite number. |
{
"input": {
"firstOutput": 0.2,
"secondOutput": 0.8
},
"output": {
"logOddsContrast": 2.7725887222397816
}
}Limitations and assumptions
- These are natural-log odds, not probability differences or base-ten odds. Values are supplied rather than model-verified; exact zero/one are rejected, and no model fit, calibration or causal meaning is inferred.
- Inputs must satisfy both the field types and the cross-field conditions above. Invalid input causes an error; a successful false flag or documented placeholder is a different result.
- Numeric calculations use binary floating point unless exact decimal-integer arithmetic is explicitly stated. No hidden measurement tolerance or decimal rounding is added; nonfinite or out-of-contract results are rejected. Numeric signed-zero results are exported as zero.
Alternative or Complementary methods
Use the complete model, link function and calibration or statistical analysis for inference. Check output boundaries and numerical precision before interpreting local log-odds arithmetic as an effect, odds ratio estimate or verified model behavior.
References
Similar or alternative tools
- activation/sigmoid
Apply the logistic sigmoid activation.
- activation/softmax-pairwise-logit-odds
Return (firstLogit-secondLogit)/temperature as the natural-log ratio of their softmax probabilities.