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

Long Form Question Answering on ELI5

31.15ROUGE-L

InSemRAG

Updated 1d ago

Evaluation Results

Method
InSemRAG 2026.05	31.15	35.42	13.45	-	-
InSemRAG 2026.05	30.85	34.85	13.15	-	-
SELF-RAG 2026.05	30.2	34.09	12.83	-	-
SELF-RAG 2026.05	29.6	33.41	12.57	-	-
InSemRAG 2026.05	28.95	32.85	12.45	-	-
SELF-RAG 2026.05	27.78	31.36	11.8	-	-
FiD 2026.05	27.37	34.72	10.52	-	-
IRCoT 2026.05	27.34	31.28	9.63	-	-
Wikipedia Graph 2026.05	26.91	32.01	9.15	-	-
FiD 2026.05	26.82	34.03	10.31	-	-
IRCoT 2026.05	26.79	30.65	9.44	-	-
Wikipedia Graph 2026.05	26.37	31.37	8.97	-	-
SearChain 2023.04	25.57	-	-	-	-
CoT-on-Concat 2026.05	25.31	31.2	8.39	-	-
FiD 2026.05	25.18	31.94	9.68	-	-
IRCoT 2026.05	25.15	28.78	8.86	-	-
SearChain w/o Completion 2023.04	25.02	-	-	-	-
ReAct 2026.05	24.81	30.21	7.26	-	-
CoT-on-Concat 2026.05	24.8	30.58	8.22	-	-
Wikipedia Graph 2026.05	24.76	29.45	8.42	-	-
Plan-and-Solve w/ IR 2023.04	24.56	-	-	-	-
Naïve RAG 2026.05	24.32	30.21	6.77	-	-
ReAct 2026.05	24.31	29.61	7.11	-	-
Tree-of-Thought w/ IR 2023.04	24.2	-	-	-	-
React -> CoT-SC 2023.04	24.05	-	-	-	-
Naïve RAG 2026.05	23.83	29.61	6.63	-	-
Verify-and-Edit 2023.04	23.8	-	-	-	-
Multi-Hop RAG 2026.05	23.61	27.86	6.32	-	-
DSP 2023.04	23.46	-	-	-	-
Direct Retrieval 2023.04	23.4	-	-	-	-
CoT-on-Concat 2026.05	23.29	28.7	7.72	-	-
Self-Ask 2023.04	23.25	-	-	-	-
Multi-Hop RAG 2026.05	23.14	27.3	6.19	-	-
ToolFormer 2023.04	23.05	-	-	-	-
SearChain w/o Verification 2023.04	22.98	-	-	-	-
ReAct 2026.05	22.83	27.79	6.68	-	-
SearChain w/o IR 2023.04	22.54	-	-	-	-
Naïve RAG 2026.05	22.37	27.79	6.23	-	-
Plan-and-Solve 2023.04	22.23	-	-	-	-
Recite-and-answer 2023.04	22.1	-	-	-	-
COT-SC 2023.04	22.05	-	-	-	-
Least-to-Most 2023.04	21.95	-	-	-	-
Direct Prompting 2023.04	21.9	-	-	-	-
CoT 2023.04	21.79	-	-	-	-
Self-Ask w/o IR 2023.04	21.73	-	-	-	-
Multi-Hop RAG 2026.05	21.72	25.63	5.81	-	-
Auto-CoT 2023.04	21.55	-	-	-	-
w/o RAG 2026.05	20.51	23.62	4.25	-	-
w/o RAG 2026.05	20.1	23.15	4.16	-	-
w/o RAG 2026.05	18.87	21.73	3.91	-	-
Interleave 2019.11	14.63	23.32	4.79	30.8	-
Sequential 2019.11	14.52	23.02	4.8	30.63	-
Interleave 2019.11	14.41	23.13	4.62	36.1	-
Multitask 2019.11	14.35	22.48	4.63	31.66	-
Concatenate 2019.11	14.33	22.86	4.54	36.37	-
Alternate 2019.11	14.24	22.79	4.54	36.95	-
Sequential 2019.11	13.83	22.14	4.24	50.23	-
Qwen3-8B 2026.04	-	-	-	-	57.1
Standard RAG 2026.04	-	-	-	-	61.5
SelfRAG 2026.04	-	-	-	-	67.4
RQ-RAG 2026.04	-	-	-	-	68.7
SOLAR 2026.04	-	-	-	-	70.2
DA-RAG 2026.04	-	-	-	-	66.2
FLARE 2026.04	-	-	-	-	69.8
DRAGIN 2026.04	-	-	-	-	70.1
P-RAG 2026.04	-	-	-	-	72.1
GuarantRAG 2026.04	-	-	-	-	74.6
Qwen3-14B 2026.04	-	-	-	-	59.2
Standard RAG 2026.04	-	-	-	-	64.3
SelfRAG 2026.04	-	-	-	-	70.1
RQ-RAG 2026.04	-	-	-	-	71.8
SOLAR 2026.04	-	-	-	-	73.2
DA-RAG 2026.04	-	-	-	-	69
FLARE 2026.04	-	-	-	-	72.7
DRAGIN 2026.04	-	-	-	-	73.1
P-RAG 2026.04	-	-	-	-	75.8
GuarantRAG 2026.04	-	-	-	-	78.3
Qwen3-14B-T 2026.04	-	-	-	-	57.4
Standard RAG 2026.04	-	-	-	-	61.8
SelfRAG 2026.04	-	-	-	-	67.2
RQ-RAG 2026.04	-	-	-	-	68.9
SOLAR 2026.04	-	-	-	-	70.6
DA-RAG 2026.04	-	-	-	-	66.1
FLARE 2026.04	-	-	-	-	69.8
DRAGIN 2026.04	-	-	-	-	70.2
P-RAG 2026.04	-	-	-	-	72.5
GuarantRAG 2026.04	-	-	-	-	74.8
GPT-4o 2026.04	-	-	-	-	76.9
o1 2026.04	-	-	-	-	78.4