Share your thoughts, 1 month free Claude Pro on usSee more

Instruction Following on AlpacaEval 2

75.4LC (%)

ξ-DPO

Updated 18d ago

Evaluation Results

Method
ξ-DPO 2026.05	75.4	67.5
SimUAPO 2025.09	73.5	67
AlphaDPO 2026.05	73.4	66.1
SimPO 2025.09	72.6	66.1
SimPO 2026.05	72.4	65
SimPO 2025.09	72.4	65.9
DPO 2026.05	70.4	66.9
UAPO 2025.09	70.2	67.7
DPO 2025.09	68.8	64.5
R-DPO 2026.05	68.3	66.9
R-DPO 2025.09	68.3	66.9
DPO 2025.09	67.8	65.4
Fine-tuned 2026.05	65.4	-
Fine-tuned 2026.05	63.6	-
IPO 2026.05	62.6	58.4
IPO 2025.09	62.6	58.4
KTO 2026.05	61.7	55.5
KTO 2025.09	61.7	55.5
DiDi-Merg.-L 2026.05	59.5	-
AlphaDPO 2026.05	58.7	51.1
DiDi-Merg.-L 2026.05	58.7	-
ξ-DPO 2026.05	57.5	50.5
CPO 2026.05	56.4	53.4
CPO 2025.09	56.4	53.4
ORPO 2026.05	56.2	46.7
ORPO 2025.09	56.2	46.7
SimPO 2026.05	55.6	49.6
FREE-Merging 2026.05	54.1	-
SimPO 2025.09	53.7	47.5
FREE-Merging 2026.05	53.6	-
Twin-Merging 2026.05	53.3	-
Twin-Merging 2026.05	52.4	-
DPO 2026.05	51.9	50.8
R-DPO 2026.05	51.6	50.7
SimUAPO 2025.09	51.2	47.6
SimPO 2025.09	50.1	45.2
RMiPO 2026.04	49.1	46.3
SFT 2026.05	48.7	36.5
SFT 2025.09	48.7	36.5
SimPER 2026.04	48.5	45.7
DPO 2025.09	48.2	47.5
R-DPO 2025.09	48	45.8
ξ-DPO 2026.05	47.1	39.1
IPO 2025.09	46.8	42.4
α-DPO 2026.04	46.6	38.1
AlphaDPO 2026.05	46.6	38.1
ϵ-DPO 2026.04	46.4	44.9
SimPO 2026.04	44.7	40.5
SimPO 2026.05	43.8	38
β-DPO 2026.04	43.4	38.2
KTO 2026.05	41.4	36.4
DPO 2025.09	41.2	37.5
UAPO 2025.09	41.2	38.3
IPO 2026.05	40.6	39.6
DPO 2026.04	40.3	37.9
R-DPO 2026.05	40.3	37.3
DPO 2026.05	40.2	38.1
KTO 2026.05	38.3	34.1
RMiPO 2026.04	38.1	37
ORPO 2025.09	38.1	33.8
Qwen2.5-14B-IT + SELF-REDTEAM 2025.06	38.09	-
SimPER 2026.04	37.8	39.5
Qwen2.5-14B-IT 2025.06	36.91	-
CPO 2026.05	36.5	40.8
ORPO 2026.05	36.5	33.1
IPO 2026.05	35.9	34.4
IPO 2026.04	35.6	35.6
ϵ-DPO 2026.04	35.6	29.6
Qwen2.5-7B-IT + SELF-REDTEAM 2025.06	34.93	-
Zero-shot 2026.05	34.2	-
CPO 2025.09	34.1	36.4
KTO 2025.09	34.1	32.1
OLMO2 INST. 2025.06	33.5	-
LLAMA3.1 (Proxy Tuned) 2025.06	33.5	-
Qwen2.5-7B-IT 2025.06	33.43	-
ξ-DPO 2026.05	33.2	36.5
KTO 2026.04	33.1	31.8
α-DPO 2026.04	32.3	32.6
AlphaDPO 2026.05	32.3	32.6
SimPO 2026.04	32.1	34.8
SimPO 2025.09	32.1	34.8
ORPO 2026.05	31.6	29.8
SimPO 2026.05	30.2	32.1
CPO 2026.05	29.6	34.4
SimPO 2025.09	29	31.9
CPO 2026.04	28.9	32.2
SimUAPO 2025.09	28.6	32.2
Zero-shot 2026.05	28.3	-
R-DPO 2025.09	27.3	24.5
SLiC 2026.04	26.9	27.5
DPO 2026.04	26.8	24.9
DPO 2025.09	26.8	24.9
CPO 2026.05	26.2	31.7
Initial 2026.04	26	25.3
ϵ-DPO 2026.04	26	23.4
SFT 2025.09	26	25.3
SimPER 2026.04	25.2	22.9
RMiPO 2026.04	25.2	23.6
Llama3.1-8B-IT 2025.06	24.74	-
α-DPO 2026.04	24.6	22.7

Showing 100 of 147 rows