Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Knowledge Attribution Causal Ablation on G-MMLU (test)
Loading...
47.5
Ablation Success Rate
XICI
3.092
14.621
26.15
37.679
Mar 17, 2026
Ablation Success Rate
Spurious Gain Rate
Rate Difference
Num. Inconsistent Qs
Num. Languages
Num. Blacklisted Experts
Num. Identified Experts Qs
Avg. Experts ID'd per Q
Orig. Correct Answers
Num. All Incorrect Qs
Updated 1mo ago
Evaluation Results
Method
Method
Links
Ablation Success Rate
Spurious Gain Rate
Rate Difference
Num. Inconsistent Qs
Num. Languages
Num. Blacklisted Experts
Num. Identified Experts Qs
Avg. Experts ID'd per Q
Orig. Correct Answers
Num. All Incorrect Qs
XICI
Model=GLM-4.5-Air, Det...
2026.03
47.5
3.4
44.1
122
42
285
122
24
1,322
17
XICI
Model=Qwen3-30B-A3B-In...
2026.03
37.6
3.4
34.2
124
42
240
124
22.8
1,017
14
Random Question-Shuffling (baseline)
Model=GLM-4.5-Air, Det...
2026.03
8
4
4
-
-
-
-
-
-
-
Random Question-Shuffling (baseline)
Model=Qwen3-30B-A3B-In...
2026.03
7.6
4.2
3.4
-
-
-
-
-
-
-
Random Expert Set of Same Size (baseline)
Model=Qwen3-30B-A3B-In...
2026.03
5.1
3.6
1.5
-
-
-
-
-
-
-
Random Expert Set of Same Size (baseline)
Model=GLM-4.5-Air, Det...
2026.03
4.8
3.7
1
-
-
-
-
-
-
-
Feedback
Search any
task
Search any
task