List your company
ResearchFormerly WizWand

Topic Classification on DBPedia

99.1AccuracyFADS-ICL
74.55680.92887.393.672Dec 2, 2022Jun 30, 2023Jan 27, 2024Aug 25, 2024Mar 24, 2025Oct 21, 2025May 20, 2026
Updated 5mo ago

Evaluation results

MethodLinks
FADS-ICLBackbone=Llama-2 7B, Shots=128-shots2024.0599.1
kNN-promptingBackbone=Llama-1 13B, Shots=128-shots2024.0599.1
kNN-promptingBackbone=Llama-1 7B, Shots=128-shots2024.0599
kNN-promptingBackbone=Llama-2 7B, Shots=128-shots2024.0599
FADS-ICLBackbone=Llama-1 13B, Shots=128-shots2024.0599
FADS-ICLBackbone=Llama-1 7B, Shots=128-shots2024.0598.9
FADS-ICLBackbone=Llama-1 30B, Shots=128-shots2024.0598.9
FADS-ICLBackbone=Llama-2 70B, Shots=128-shots2024.0598.9
kNN-promptingBackbone=Llama-2 13B, Shots=128-shots2024.0598.8
FADS-ICLBackbone=Llama-2 13B, Shots=128-shots2024.0598.8
kNN-promptingBackbone=Llama-1 30B, Shots=128-shots2024.0598.8
kNN-promptingBackbone=Llama-2 70B, Shots=128-shots2024.0598.8
FADS-ICLcandidate pool size (m)=2562024.0598.7
FADS-ICLNumber of training shots (m)=128, LLM scale=1.5B2024.0598.6
FADS-ICLBackbone=GPT-2 1.5B, Shots=128-shots2024.0598.6
FADS-ICLcandidate pool size (m)=1282024.0598.6
kNN-promptBackbone=Llama-2 7B, Shots=128-shots2024.0598.3
FADS-ICLNumber of training shots (m)=64, LLM scale=1.5B2024.0598.1
FADS-ICLcandidate pool size (m)=642024.0598.1
FADS-ICLcandidate pool size (m)=322024.0597.9
FADS-ICLBackbone=GPT-2 0.8B, Shots=128-shots2024.0597.7
ICLModel=LLaMA-3.1-8B2026.0597.68
kNN-promptBackbone=Llama-2 13B, Shots=128-shots2024.0597.3
FADS-ICLNumber of training shots (m)=16, LLM scale=1.5B2024.0597.2
ICLModel=LLaMA-2-13B2026.0596.88
kNN-promptingNumber of training shots (m)=128, LLM scale=1.5B2024.0596.8
kNN-promptingBackbone=GPT-2 1.5B, Shots=128-shots2024.0596.8
GPT Large FTBackbone=GPT Large, Shots=N/A (Fine-tuned)2024.0596.5
ICLBackbone=Llama-2 7B, Shots=128-shots2024.0595.7
kNN-promptingBackbone=GPT-2 0.8B, Shots=128-shots2024.0595.5
FADS-ICLNumber of training shots (m)=8, LLM scale=1.5B2024.0595.3
kNN-promptingNumber of training shots (m)=64, LLM scale=1.5B2024.0595.2
BERT Large FTBackbone=BERT Large, Shots=N/A (Fine-tuned)2024.0595.1
kNN-promptBackbone=Llama-1 13B, Shots=128-shots2024.0595.1
kNN-promptBackbone=Llama-1 30B, Shots=128-shots2024.0595.1
DecTn (Shots)=162022.1294.6
ICLBackbone=Llama-2 70B, Shots=128-shots2024.0594.1
ICLBackbone=Llama-1 13B, Shots=128-shots2024.0593.8
kNN-promptingNumber of training shots (m)=16, LLM scale=1.5B2024.0593.5
LTVModel=LLaMA-3.1-8B2026.0593.36
ICLBackbone=Llama-1 7B, Shots=128-shots2024.0593.3
LTVModel=LLaMA-2-13B2026.0593.2
SubclusteringAnnotation budget (|L|)=1002024.0692.32
kNN-promptBackbone=Llama-1 7B, Shots=128-shots2024.0591.6
ICLBackbone=Llama-1 30B, Shots=128-shots2024.0591.2
kNN-promptingNumber of training shots (m)=8, LLM scale=1.5B2024.0589.9
Trans-Encodercandidate pool size (m)=2562024.0589.9
kNN-promptBackbone=Llama-2 70B, Shots=128-shots2024.0589.5
State VectorModel=LLaMA-2-13B2026.0589.48
SimCSEcandidate pool size (m)=2562024.0589.1
SubclusteringAnnotation budget (|L|)=182024.0689.06
FastGASModel=OPT-6.7B, Annotation Budget=182024.0688.93
SBERTcandidate pool size (m)=2562024.0588.5
BM25candidate pool size (m)=2562024.0588.2
Vote-kModel=OPT-6.7B, Annotation Budget=182024.0688.02
SBERTcandidate pool size (m)=1282024.0587.2
SimCSEcandidate pool size (m)=1282024.0587
Trans-Encodercandidate pool size (m)=1282024.0587
kNN-promptNumber of training shots (m)=64, LLM scale=1.5B2024.0586.4
kNN-promptNumber of training shots (m)=128, LLM scale=1.5B2024.0586.4
kNN-promptBackbone=GPT-2 1.5B, Shots=128-shots2024.0586.4
kNN-promptNumber of training shots (m)=16, LLM scale=1.5B2024.0585.8
kNN-promptNumber of training shots (m)=32, LLM scale=1.5B2024.0585.7
Trans-Encodercandidate pool size (m)=642024.0585.6
BM25candidate pool size (m)=1282024.0585.5
kNN-promptNumber of training shots (m)=4, LLM scale=1.5B2024.0585.2
MUPPET2022.1285.17
kNN-promptNumber of training shots (m)=8, LLM scale=1.5B2024.0584.9
SimCSEcandidate pool size (m)=642024.0584.8
SBERTcandidate pool size (m)=642024.0584.6
SBERTcandidate pool size (m)=322024.0583.8
In-Context Learning2024.0583.5
BM25candidate pool size (m)=642024.0583.4
Task VectorModel=LLaMA-3.1-8B2026.0583.24
IDEALModel=OPT-6.7B, Annotation Budget=182024.0683.2
Trans-Encodercandidate pool size (m)=322024.0582.8
SimCSEcandidate pool size (m)=322024.0582.3
ICLNumber of training shots (m)=4, LLM scale=1.5B2024.0582
ICLNumber of training shots (m)=8, LLM scale=1.5B2024.0582
ICLNumber of training shots (m)=16, LLM scale=1.5B2024.0582
ICLNumber of training shots (m)=32, LLM scale=1.5B2024.0582
ICLNumber of training shots (m)=64, LLM scale=1.5B2024.0582
ICLNumber of training shots (m)=128, LLM scale=1.5B2024.0582
ICLBackbone=GPT-2 1.5B, Shots=128-shots2024.0582
ICLBackbone=Llama-2 13B, Shots=128-shots2024.0581.5
FastGASModel=GPT-Neo-2.7B, Annotation Budget=182024.0680.86
State VectorModel=LLaMA-3.1-8B2026.0580.8
Vote-kModel=GPT-Neo-2.7B, Annotation Budget=182024.0680.73
Task VectorModel=LLaMA-2-13B2026.0580.72
BM25candidate pool size (m)=322024.0580.7
FADS-ICLNumber of training shots (m)=32, LLM scale=1.5B2024.0579.9
kNN-promptingNumber of training shots (m)=4, LLM scale=1.5B2024.0579.1
I2CLModel=LLaMA-2-13B2026.0579
IDEALModel=GPT-Neo-2.7B, Annotation Budget=182024.0678.38
ColD-Fusion2022.1278.15
Multitask2022.1277.69
Finetune2022.1277.49
Function VectorModel=LLaMA-2-13B2026.0576.4
Zero-shotModel=LLaMA-2-13B2026.0576.2
Latent Concept LearningLLM=GPT3-d (175B)2023.0175.5