ResearchBenchmarksList operations evaluation on ListOps (3, 9) (test)Follow89.9Mean AccuracyRIR-GRC33.7448.3262.977.48May 25, 2026Evaluation ResultsMethodMethodLinksMean AccuracyLongest Bin PerformanceRIR-GRC2026.0589.979.2BBT-GRC2026.0584.172.5MLP-LDRU2026.0574.767.8TF (ALiBi)Positional Encoding=ALiBiPositional Encoding=ALiBi2026.0568.255.1LSTM2026.0565.950TF (NoPE)Positional Encoding=NonePositional Encoding=None2026.0537.332.1TF (Sin.)Positional Encoding=Si...Positional Encoding=Sinusoidal2026.0535.99.8