Scholarly Abstract Generation on LaMP Scholar benchmark leaderboard · SOTA2 Research

Scholarly Abstract Generation on LaMP Scholar

44.6ROUGE-1RankGPT (GPT-5-nano)

Evaluation Results

Method	Links
RankGPT (GPT-5-nano)2026.01		44.6	38.8	41.5
PURPLE2026.01		44.4	38.8	41
ICR (Llama-3-8B-Instruct)2026.01		44.3	38.5	41.2
Contriever2026.01		44.1	38.6	41.1
RankGPT (Llama-3-8B-Instruct)2026.01		44	38.6	41.1
BM252026.01		43.8	38.2	40.5
REPLUG2026.01		42.5	37.1	40.8
REPLUG2026.01		41.8	36.6	38.6
PURPLE2026.01		41.4	35.8	40.3
ICR (Llama-3-8B-Instruct)2026.01		41.3	35.5	40.8
Contriever2026.01		41.1	35.4	40.5
BM252026.01		41	35.2	40.6
RankGPT (Llama-3-8B-Instruct)2026.01		41	35.4	40.5
RankGPT (GPT-5-nano)2026.01		41	35.3	40.5
PURPLE2026.01		40	33.5	42.4
RankGPT (GPT-5-nano)2026.01		39.8	33.4	42.3
BM252026.01		39.7	33.2	42.2
Contriever2026.01		39.7	33.3	42
RankGPT (Llama-3-8B-Instruct)2026.01		39.7	33.3	42
IC-RALM2026.01		39.7	34.1	38.6
ICR (Llama-3-8B-Instruct)2026.01		39.5	33.1	42
REPLUG2026.01		38.1	32.8	40.4
IC-RALM2026.01		37.6	31	41
IC-RALM2026.01		37.5	31.5	39.2