Preference Alignment on Anthropic-hh-rlhf (test) benchmark leaderboard · SOTA2 Research

Preference Alignment on Anthropic-hh-rlhf (test)

5.83LLM-as-a-Judge Helpful ScorePLC

Evaluation Results

Method	Links
PLC2026.04		5.83	9.29	7.56	1.08	0.09	0.61
SOLO-R12026.04		5.79	8.47	7.13	1.91	1.23	0.44
PLC2026.04		5.71	9.28	7.49	1.93	0.18	0.89
SOLO-R12026.04		5.68	8.1	6.89	1.04	0.89	0.33
GAPO2026.04		5.3	8.83	7.07	0.8	0.18	0.31
GAPO2026.04		5.18	8.87	7.03	1.72	0.33	0.75
Ric2026.04		4.46	8.68	6.57	0.85	0.17	0.59
Ric2026.04		4.41	8.56	6.49	1.42	0.1	0.66
RS2026.04		4.03	8.33	6.18	0.68	0.09	0.29
RS2026.04		3.92	8.25	6.09	1.2	0.08	0.56
SOLO-R02026.04		3.57	8.71	6.13	0.35	1.12	0.39
SOLO-R02026.04		3.54	8.63	6.09	0.35	0.54	0.09