List your company

Classification Probing on Sarcasm (test)

96.3Probe Acc (Best Layer)xRFM
87.66889.90992.1594.391Jan 29, 2026

Evaluation results

MethodLinks
xRFMBackbone=Llama-3.1-70B-Instruct2026.0196.392.9
RAPTORBackbone=Llama-3.1-70B-Instruct2026.0196.393
xRFMBackbone=Llama-3.3-70B-Instruct2026.0195.491.8
RAPTORBackbone=Llama-3.3-70B-Instruct2026.019592.2
xRFMBackbone=Qwen-32B-Instruct2026.0194.789.5
RAPTORBackbone=Qwen-32B-Instruct2026.0194.389.5
RAPTORBackbone=Llama-3.1-8B-Instruct2026.0194.390.1
GCSBackbone=Llama-3.1-70B-Instruct2026.0194.290.6
xRFMBackbone=Llama-3.1-8B-Instruct2026.019490.2
GCSBackbone=Llama-3.3-70B-Instruct2026.0193.689.8
RAPTORBackbone=Qwen-7B-Instruct2026.0192.687.7
xRFMBackbone=Qwen-7B-Instruct2026.0192.587.5
GCSBackbone=Qwen-32B-Instruct2026.0192.387.1
xRFMBackbone=Qwen-3B-Instruct2026.0191.985.9
GCSBackbone=Llama-3.1-8B-Instruct2026.0191.787.6
RAPTORBackbone=Qwen-3B-Instruct2026.0191.585.7
GCSBackbone=Qwen-7B-Instruct2026.0190.285
RAPTORBackbone=Gemma-7B-it2026.0190.285.2
xRFMBackbone=Gemma-7B-it2026.0189.584.9
GCSBackbone=Qwen-3B-Instruct2026.0189.482
GCSBackbone=Gemma-7B-it2026.018882