Our new X account is live! Follow @wizwand_team for updates

General Capability Evaluation on tinyBenchmarks

90AI2_arc Accuracy

AAS

Updated 4d ago

Evaluation Results

Method	Links
AAS 2026.01		90	91	69	47	45
No Steering 2026.01		90	93	71	74	59
DirAbl 2026.01		90	94	70	74	59
AAS 2026.01		90	93	72	75	57
Selective Steering 2026.01		90	93	71	74	59
No Steering 2026.01		87	93	64	63	59
Selective Steering 2026.01		87	94	65	63	59
DirAbl 2026.01		86	92	64	57	61
No Steering 2026.01		81	85	66	56	51
DirAbl 2026.01		80	86	67	56	49
Selective Steering 2026.01		80	88	66	55	51
No Steering 2026.01		80	88	61	60	53
DirAbl 2026.01		80	82	62	57	50
Selective Steering 2026.01		79	88	61	61	53
ActAdd 2026.01		79	81	68	36	49
SAS 2026.01		79	81	68	36	49
AAS 2026.01		78	85	52	34	50
AAS 2026.01		77	88	67	57	47
DirAbl 2026.01		73	65	56	58	43
Selective Steering 2026.01		72	78	61	57	37
No Steering 2026.01		71	80	61	57	36
No Steering 2026.01		71	70	54	55	38
Selective Steering 2026.01		71	69	54	56	40
AAS 2026.01		70	81	59	56	42
AAS 2026.01		70	72	50	51	45
No Steering 2026.01		69	78	53	49	47
Selective Steering 2026.01		69	72	52	48	47
DirAbl 2026.01		67	75	61	59	34
DirAbl 2026.01		66	76	48	43	43
ActAdd 2026.01		61	64	51	39	35
SAS 2026.01		61	64	51	39	35
No Steering 2026.01		47	43	46	21	31
Selective Steering 2026.01		46	46	42	22	31
AAS 2026.01		45	35	42	20	36
DirAbl 2026.01		41	40	38	31	35
ActAdd 2026.01		41	68	33	39	36
AAS 2026.01		38	8	13	14	27
SAS 2026.01		24	46	15	20	29
ActAdd 2026.01		17	12	7	3	2
SAS 2026.01		17	12	7	3	2
ActAdd 2026.01		8	0	6	18	10
SAS 2026.01		8	0	8	37	17
ActAdd 2026.01		1	0	0	0	0
SAS 2026.01		1	0	0	0	0
ActAdd 2026.01		0	0	0	1	0
SAS 2026.01		0	0	0	1	0
ActAdd 2026.01		0	0	0	0	0
SAS 2026.01		0	0	0	0	0