Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Opinion Alignment on American National Election Studies (ANES) 2020 Time Series (test)
Loading...
45.43
Mean Macro-F1
SFT+GRPO
18.182
25.256
32.33
39.404
Mar 1, 2026
Mean Macro-F1
Updated 3mo ago
Evaluation Results
Method
Method
Links
Mean Macro-F1
SFT+GRPO
Base model=Magistral 24B
2026.03
45.43
GRPO
Base model=Magistral 24B
2026.03
43.79
SFT
Base model=Llama 3.1 8B
2026.03
42.77
SFT+GRPO
Base model=Llama 3.1 8B
2026.03
40.66
SFT
Base model=Magistral 24B
2026.03
39.15
SFT+GRPO
Base model=Qwen3 8B
2026.03
38.44
SFT
Base model=Qwen3 8B
2026.03
35.14
ORPO
Base model=Llama 3.1 8B
2026.03
34.84
GRPO
Base model=Llama 3.1 8B
2026.03
34.55
random
Base model=Untrained b...
2026.03
33.33
GRPO
Base model=Qwen3 8B
2026.03
31.47
ORPO
Base model=Qwen3 8B
2026.03
26.95
ORPO
Base model=Magistral 24B
2026.03
24.25
icl
Base model=Llama 3.1 8B
2026.03
23.2
icl
Base model=Qwen3 8B
2026.03
23.2
majority
Base model=Untrained b...
2026.03
22.98
icl
Base model=Magistral 24B
2026.03
19.23
Feedback
Search any
task
Search any
task