List your company

Classification Probing on STSA (test)

0.955Probe Accuracy (Best Layer)RAPTOR
0.919640.928820.9380.94718Jan 29, 2026

Evaluation results

MethodLinks
RAPTORBackbone=Qwen-32B-Instruct2026.010.9550.906
xRFMBackbone=Llama-3.1-70B-Instruct2026.010.9550.923
xRFMBackbone=Qwen-32B-Instruct2026.010.9530.903
RAPTORBackbone=Llama-3.3-70B-Instruct2026.010.9520.921
RAPTORBackbone=Qwen-3B-Instruct2026.010.950.882
xRFMBackbone=Llama-3.3-70B-Instruct2026.010.950.919
xRFMBackbone=Qwen-7B-Instruct2026.010.9490.9
RAPTORBackbone=Llama-3.1-70B-Instruct2026.010.9490.924
RAPTORBackbone=Qwen-7B-Instruct2026.010.9470.902
xRFMBackbone=Llama-3.1-8B-Instruct2026.010.9470.916
RAPTORBackbone=Llama-3.1-8B-Instruct2026.010.9470.918
GCSBackbone=Qwen-32B-Instruct2026.010.9450.883
xRFMBackbone=Qwen-3B-Instruct2026.010.9430.877
GCSBackbone=Llama-3.1-8B-Instruct2026.010.9390.899
GCSBackbone=Llama-3.3-70B-Instruct2026.010.9390.898
GCSBackbone=Qwen-7B-Instruct2026.010.9380.877
GCSBackbone=Llama-3.1-70B-Instruct2026.010.9380.902
GCSBackbone=Qwen-3B-Instruct2026.010.9350.836
xRFMBackbone=Gemma-7B-it2026.010.9320.886
RAPTORBackbone=Gemma-7B-it2026.010.930.889
GCSBackbone=Gemma-7B-it2026.010.9210.856