Value Alignment on cross-demographic (test)
48.6AccuracyQwen3-8B-DVMap
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3-8B-DVMapLLM=Qwen3-8B-DVMap2026.05 | 48.6 | 83.9 | 0.1321 | |
| Gpt-4oLLM=Gpt-4o2026.05 | 48.5 | 83.8 | 0.1418 | |
| Qwen3-next-80B-a3BLLM=Qwen3-next-80B-a3B2026.05 | 47.6 | 82.5 | 0.1449 | |
| Llama-3.3-70B-InstructLLM=Llama-3.3-70B-Instruct2026.05 | 46.4 | 83.3 | 0.1504 | |
| Gpt-4o-miniLLM=Gpt-4o-mini2026.05 | 46.3 | 82.4 | 0.1476 | |
| Qwen3-14BLLM=Qwen3-14B2026.05 | 46.2 | 83.5 | 0.146 | |
| Gemini-2.5-flash-liteLLM=Gemini-2.5-flash-lite2026.05 | 45.3 | 79.7 | 0.1538 | |
| DeepSeek-v3.2-expLLM=DeepSeek-v3.2-exp2026.05 | 45.1 | 82.3 | 0.1342 | |
| Claude-3.7-sonnetLLM=Claude-3.7-sonnet2026.05 | 26.9 | 46.4 | 0.1503 |