ECHO: Efficient Causal Self-Recursion Beyond Auxiliary Multi-Token Prediction
Open paper
Benchmark
Model
Rank
Results
question-answering-on-piqa
ECHO-1.5(3)
#76
Accuracy: 78.38
common-sense-reasoning-on-winogrande
ECHO-1.5(3)
#95
Accuracy: 64.4
Rank counts only results with a code link.