Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multiple Choice Question Answering on MMLU (Performance Change Tracking)
Loading...
58
MMLU Accuracy Gain (Absolute)
FLOATDOOR
-1.3632
14.0484
29.46
44.8716
Feb 11, 2026
Mar 4, 2026
Mar 25, 2026
Apr 15, 2026
May 6, 2026
May 27, 2026
Jun 17, 2026
MMLU Accuracy Gain (Absolute)
MMLU Baseline Accuracy (Before)
MMLU Post-Intervention Accuracy (After)
Updated 1mo ago
Evaluation Results
Method
Method
Links
MMLU Accuracy Gain (Absolute)
MMLU Baseline Accuracy (Before)
MMLU Post-Intervention Accuracy (After)
FLOATDOOR
Scenario=M1
2026.06
58
-
-
FLOATDOOR
Scenario=M2
2026.06
35
-
-
FLOATDOOR
Scenario=S3
2026.06
28
-
-
FLOATDOOR
Scenario=S2
2026.06
7
-
-
FLOATDOOR
Scenario=S1
2026.06
2
-
-
CRL-Token
Backbone=LLaMA-3.1 8B,...
2026.02
0.92
-
62.33
LLaMA-3.1 8B
Evaluation Protocol=0-...
2026.02
-
61.41
-
Feedback
Search any
task
Search any
task