Description
For the requested claim, count distinct source IDs with true and false support separately.
Evidence weights and credences are numerical representations under a chosen model. Likelihood ratios, probabilities, logarithmic weights and confidence bounds have different meanings; one local calculation does not establish the truth of a claim.
For the requested claim, count distinct source IDs with true and false support separately.
When to use this tool
Use Claim Source Support Count when you need an explicit evidence-weight or probability-consistency term.
How to use the tool and how it works
- Enter claim, claimTokens, sourceIds, supports. Keep parallel arrays aligned. Use the units and conventions below.
- Run the tool. Correct any format, range, or cross-field input error before interpreting the output.
- Read supportingSources, opposingSources. Use the output convention below.
Require aligned claim tokens, source IDs and support flags. For the requested claim, count distinct source IDs with true and false support separately.1
| Input field | Meaning and units | Accepted input |
|---|---|---|
| claim | Claim | Required. Text; maximum UTF-16 length 8192. |
| claimTokens | Claim Tokens | Required. Array of text; maximum UTF-16 length 8192; maximum items 256. |
| sourceIds | Source Ids | Required. Array of text; maximum UTF-16 length 8192; maximum items 256. |
| supports | Supports | Required. Array of boolean; maximum items 256. |
| Output field | Interpretation | Returned shape |
|---|---|---|
| supportingSources | Supporting Sources | Required. Safe integer. |
| opposingSources | Distinct IDs with a false support record under the supplied binary convention; may overlap supportingSources. | Required. Safe integer. |
{
"input": {
"claim": "A",
"claimTokens": [
"A",
"A",
"B"
],
"sourceIds": [
"s1",
"s2",
"s3"
],
"supports": [
true,
false,
true
]
},
"output": {
"supportingSources": 1,
"opposingSources": 1
}
}Limitations and assumptions
- One source can appear in both sets through conflicting records. False means opposing under this binary convention, not missing or neutral evidence; sources are not authenticated.
- 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.
Alternative or Complementary methods
Evaluate source quality and the complete inference model. Check likelihood meaning, dependence, prior assumptions, normalization and uncertainty rather than substituting counts or weights for justified evidence.
References
Similar or alternative tools
- Testimonial Source Dependence Discount Term
Multiply logEvidence by weight×count and return that fractional equivalent source count.
- Assertion Support Path Count
Count distinct selected-assertion paths and any-valid paths; subtract for invalidOnlyPaths.