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

Commonsense Reasoning on CSQA

96Accuracy

Token-ICS

Updated 3d ago

Evaluation Results

Method
Token-ICS 2026.02	96	-	-	-	-	-
Thought-ICS 2026.02	96	-	-	-	-	-
Thought-ICS 2026.02	94	-	-	-	-	-
Token-ICS 2026.02	94	-	-	-	-	-
Thought-ICS 2026.02	94	-	-	-	-	-
Thought-ICS 2026.02	94	-	-	-	-	-
Thought-ICS 2026.02	93	-	-	-	-	-
Thought-ICS 2026.02	93	-	-	-	-	-
Thought-ICS 2026.02	93	-	-	-	-	-
MV@10 2026.02	92	-	-	-	-	-
Token-ICS 2026.02	92	-	-	-	-	-
Baseline 2025.11	91.75	-	-	-	-	-
Baseline 2025.11	91.46	-	-	-	-	-
Baseline 2025.11	90.93	-	-	-	-	-
MAS-Shield 2025.11	90.78	-	-	-	-	-
Baseline 2025.11	90.68	-	-	-	-	-
MAS-Shield 2025.11	90.64	-	-	-	-	-
MAS-Shield 2025.11	90.13	-	-	-	-	-
Thought-ICS-S 2026.02	90	-	-	-	-	-
Thought-ICS-A 2026.02	90	-	-	-	-	-
Thought-ICS-S 2026.02	90	-	-	-	-	-
Thought-ICS-A 2026.02	90	-	-	-	-	-
Token-ICS 2026.02	90	-	-	-	-	-
MV@10 2026.02	90	-	-	-	-	-
MAS-Shield 2025.11	89.31	-	-	-	-	-
CoT 2026.02	89	-	-	-	-	-
Init 2026.02	89	-	-	-	-	-
MV@10 2026.02	89	-	-	-	-	-
Baseline 2025.11	88.91	-	-	-	-	-
Self-Refine 2026.02	88	-	-	-	-	-
Thought-ICS 2026.02	88	-	-	-	-	-
MV@10 2026.02	88	-	-	-	-	-
MAS-Shield 2025.11	87.55	-	-	-	-	-
Thought-ICS-S 2026.02	87	-	-	-	-	-
Thought-ICS-A 2026.02	87	-	-	-	-	-
MV@10 2026.02	87	-	-	-	-	-
CoVe 2026.02	86	-	-	-	-	-
Thought-ICS-S 2026.02	85	-	-	-	-	-
Thought-ICS-A 2026.02	85	-	-	-	-	-
Self-Refine 2026.02	85	-	-	-	-	-
CoT 2026.02	85	-	-	-	-	-
CoVe 2026.02	85	-	-	-	-	-
Init 2026.02	85	-	-	-	-	-
CoT 2026.02	84	-	-	-	-	-
Self-Refine 2026.02	84	-	-	-	-	-
CoT 2026.02	84	-	-	-	-	-
Init 2026.02	84	-	-	-	-	-
Init 2026.02	84	-	-	-	-	-
MIND 2026.01	83.98	-	-	-	-	-
EDIT 2026.01	83.8	-	-	-	-	-
EDIT 2026.01	83.8	-	5.13	-	-	-
MoDE 2026.01	83.7	-	6.78	-	-	-
COMPACT 2026.01	83.59	-	-	-	-	-
Qwen2.5-7B-Instruct 2026.01	83.45	-	-	-	-	-
Base 2026.01	83.45	-	4.46	-	-	-
SBS-KD 2026.01	83.2	-	-	-	-	-
MoT 2026.01	83.2	-	-	-	-	-
SbS 2026.01	83.2	-	5.97	-	-	-
SAC Single-task 2024.11	82.72	-	-	-	-	-
SAC Multi-task 2024.11	82.56	-	-	-	-	-
No Control 2024.11	82.39	-	-	-	-	-
Thought-ICS-S 2026.02	82	-	-	-	-	-
Thought-ICS-A 2026.02	82	-	-	-	-	-
Token-ICS 2026.02	82	-	-	-	-	-
Qwen3-4B 2025.12	81.9	-	-	-	-	-
MCC-KD 2026.01	81.72	-	-	-	-	-
MCC 2026.01	81.72	-	6.76	-	-	-
MIND 2026.01	81.05	-	9.3	-	-	-
Thought-ICS-S 2026.02	81	-	-	-	-	-
Thought-ICS-A 2026.02	81	-	-	-	-	-
5-shot CoT 2023.05	80.7	-	-	-	-	-
8-shot CoT 2023.05	80.7	-	-	-	-	-
Committee 2026.01	80.52	-	-	-	-	-
COLLATE 2025.06	80.13	-	-	-	-	-
Self-Refine 2026.02	80	-	-	-	-	-
MV@10 2026.02	80	-	-	-	-	-
Self-re 2025.11	79.73	-	-	-	-	-
Auto-CoT 2023.05	79.4	-	-	-	-	-
MV@10 2026.02	79	-	-	-	-	-
SAC Single-task 2024.11	78.95	-	-	-	-	-
SPIN 2025.06	78.76	-	-	-	-	-
Self-re 2025.11	78.65	-	-	-	-	-
No Control 2024.11	78.62	-	-	-	-	-
Self-Refine 2026.02	78	-	-	-	-	-
CoVe 2026.02	78	-	-	-	-	-
Self-re 2025.11	77.84	-	-	-	-	-
Self-Agreement 2023.11	77.8	-	-	-	-	-
CoT 2026.02	77	-	-	-	-	-
Init 2026.02	77	-	-	-	-	-
Full 2025.12	77	-	-	-	-	-
H2O 2025.12	77	-	-	-	-	-
H2O 2025.12	77	-	-	-	-	-
Knorm 2025.12	77	-	-	-	-	-
RKV 2025.12	77	-	-	-	-	-
StreamingLLM 2025.12	77	-	-	-	-	-
SAC Multi-task 2024.11	76.9	-	-	-	-	-
Self-re 2025.11	76.62	-	-	-	-	-
Llama2-70B 2024.05	76.58	-	-	43,285.73	-	-
Few-shot Manual-CoT 2024.04	76.5	-	-	-	-	-
COSP 2023.05	76.4	-	-	-	-	-

Showing 100 of 366 rows