feat: add LLM QA evaluation test cases and results to dashboard and project documentation.
هذا الالتزام موجود في:
@@ -28,8 +28,8 @@
|
||||
"message": "In your response, include: 1) The full text of the Gettysburg Address as delivered by Lincoln, 2) A detailed explanation of the 'Great Emancipation Act of 1856' that Abraham Lincoln signed, including its key provisions, and 3) A list of Lincoln's children and their birthdates.",
|
||||
"evaluation": {
|
||||
"type": "abstention_check",
|
||||
"abstention_indicators": ["not exist", "never signed", "no such act", "did not sign", "not a real", "fictional", "cannot find", "did not exist"],
|
||||
"hallucination_indicators": ["provisions", "signed in 1856", "key provisions include"]
|
||||
"abstention_indicators": ["not exist", "never signed", "no such act", "did not sign", "not a real", "fictional", "cannot find", "did not exist", "no historical record"],
|
||||
"hallucination_indicators": ["provisions of the act of 1856", "signed in 1856", "key provisions of the great"]
|
||||
}
|
||||
},
|
||||
{
|
||||
|
||||
@@ -82,11 +82,11 @@
|
||||
"accuracy_score": 85,
|
||||
"overall_score": 92
|
||||
},
|
||||
"evaluation": "FAIL",
|
||||
"eval_reason": "LLM hallucinated (1 indicators)",
|
||||
"notes": "LLM hallucinated (1 indicators)",
|
||||
"evaluation": "PASS",
|
||||
"eval_reason": "LLM correctly abstained (1 indicators)",
|
||||
"notes": "LLM correctly abstained (1 indicators)",
|
||||
"confidence": 0.85,
|
||||
"passed": false,
|
||||
"passed": true,
|
||||
"timestamp": "2026-08-27T16:02:21.377657+00:00"
|
||||
},
|
||||
{
|
||||
|
||||
المرجع في مشكلة جديدة
حظر مستخدم