List your company
ResearchFormerly WizWand

Topic Classification on AG-News

95.58AccuracyTay et al. (2022)
85.824888.357490.8993.4226Nov 15, 2022Jun 22, 2023Jan 28, 2024Sep 3, 2024Apr 11, 2025Nov 16, 2025Jun 24, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Tay et al. (2022)Mode=Fully-supervised, Model Parameters=4x of UD (~44B)2022.1195.58
UD+-XXLMode=Fully-supervised, Backbone=T5-XXL encoder, Model Parameters=11B2022.1195.56
Yang et al., 2019#Params (M)=-2026.0695.5
BERT2026.0494.75
RLLR_MIXEDBackbone=LLaMA2 7B2024.0593.5
RLLRBackbone=LLaMA2 7B2024.0593.4
RLLRBackbone=ChatGLM3 6B2024.0593.4
RLLR_MIXEDBackbone=ChatGLM3 6B2024.0593.4
SFTBackbone=Baichuan2 7B2024.0593.4
SFT w. rat.Backbone=ChatGLM3 6B2024.0593.1
RLHFBackbone=ChatGLM3 6B2024.0593.1
RLLRBackbone=Mistral 7B2024.0593.1
LLM-Guided TM2026.0493.1
RLHFBackbone=LLaMA2 7B2024.0593
SFTBackbone=ChatGLM3 6B2024.0593
RLLRBackbone=Baichuan2 7B2024.0593
RLLR_MIXEDBackbone=Baichuan2 7B2024.0593
RLHFBackbone=Baichuan2 7B2024.0592.9
SFT w. rat.Backbone=Mistral 7B2024.0592.7
RLLR_MIXEDBackbone=Mistral 7B2024.0592.7
RLHFBackbone=Bloom 7B2024.0592.7
RLLRBackbone=Bloom 7B2024.0592.7
GENICLBackbone=LLaMA-7B, Evaluation Setting=Few-shot (ICL)2025.0592.6
SFT w. rat.Backbone=LLaMA2 7B2024.0592.5
SFTBackbone=Mistral 7B2024.0592.5
SFT w. rat.Backbone=Baichuan2 7B2024.0592.5
RLLR_MIXEDBackbone=Bloom 7B2024.0592.5
LLM-RBackbone=LLaMA-7B, Evaluation Setting=Few-shot (ICL)2025.0592.4
SFTBackbone=LLaMA2 7B2024.0592.2
RLHFBackbone=Mistral 7B2024.0592.2
RNN/LSTMEmbedding=GloVe2026.0492.1
SFT w. rat.Backbone=Bloom 7B2024.0591.8
EPRBackbone=LLaMA-7B, Evaluation Setting=Few-shot (ICL)2025.0591.8
fastTexth=102026.0491.5
ANYSIMLITE#Params (M)=1.32026.0691.12
E5baseBackbone=LLaMA-7B, Evaluation Setting=Few-shot (ICL)2025.0590.6
FADS-ICLBackbone=Llama-1 30B, Shots=128-shots2024.0590.5
FADS-ICLBackbone=Llama-2 70B, Shots=128-shots2024.0590.2
TMEmbedding=GloVe2026.0490.12
BM25Backbone=LLaMA-7B, Evaluation Setting=Few-shot (ICL)2025.0590
kNN-promptBackbone=Llama-1 13B, Shots=128-shots2024.0589.9
kNN-promptBackbone=Llama-1 30B, Shots=128-shots2024.0589.9
Finetune2022.1289.85
SFTBackbone=Bloom 7B2024.0589.8
FADS-ICLcandidate pool size (m)=2562024.0589.8
SBERTBackbone=LLaMA-7B, Evaluation Setting=Few-shot (ICL)2025.0589.8
MUPPET2022.1289.77
FADS-ICLBackbone=Llama-1 7B, Shots=128-shots2024.0589.7
BoW TFIDF2026.0489.64
kNN-promptingBackbone=Llama-2 13B, Shots=128-shots2024.0589.6
FADS-ICLBackbone=Llama-2 13B, Shots=128-shots2024.0589.6
ColD-Fusion2022.1289.58
Multitask2022.1289.55
BERT-base + PGKD#Params (M)=-2026.0689.5
DCModel=Llama-2 13b, ICL Setting=1-shot2024.0589.34
FADS-ICLBackbone=Llama-1 13B, Shots=128-shots2024.0589.2
ICLModel=Llama-2 13b, ICL Setting=1-shot2024.0589.14
FADS-ICLBackbone=Llama-2 7B, Shots=128-shots2024.0589.1
DCModel=Llama-2 7b, ICL Setting=1-shot2024.0589.08
kNN-promptingBackbone=Llama-1 30B, Shots=128-shots2024.0589
ICLBackbone=Llama-2 70B, Shots=128-shots2024.0588.8
UniBiasModel=Llama-2 13b, ICL Setting=1-shot2024.0588.68
DCBackbone Model=Qwen, Shot Count=16-shot2025.0588.44
DCModel=Qwen, k-shot=162025.0588.44
FADS-ICLNumber of training shots (m)=128, LLM scale=1.5B2024.0588.4
FADS-ICLBackbone=GPT-2 1.5B, Shots=128-shots2024.0588.4
kNN-promptingBackbone=Llama-1 13B, Shots=128-shots2024.0588.4
FADS-ICLcandidate pool size (m)=1282024.0588.4
TM2026.0488.34
GPT Large FTBackbone=GPT Large, Shots=N/A (Fine-tuned)2024.0588.3
kNN-promptingBackbone=Llama-2 7B, Shots=128-shots2024.0588.3
UniBiasModel=Llama-2 7b, ICL Setting=1-shot2024.0588.29
CCModel=Llama-2 13b, ICL Setting=1-shot2024.0588.23
FADS-ICLNumber of training shots (m)=32, LLM scale=1.5B2024.0588.2
FADS-ICLBackbone=GPT-2 0.8B, Shots=128-shots2024.0588.1
BERT Large FTBackbone=BERT Large, Shots=N/A (Fine-tuned)2024.0588
kNN-promptBackbone=Llama-2 13B, Shots=128-shots2024.0587.9
SCBackbone Model=Mistral, Shot Count=16-shot2025.0587.81
SCModel=Mistral, k-shot=162025.0587.81
kNN-promptingNumber of training shots (m)=128, LLM scale=1.5B2024.0587.8
kNN-promptingBackbone=GPT-2 1.5B, Shots=128-shots2024.0587.8
FADS-ICLcandidate pool size (m)=642024.0587.8
kNN-promptingBackbone=Llama-2 70B, Shots=128-shots2024.0587.7
kNN-promptingBackbone=Llama-1 7B, Shots=128-shots2024.0587.6
kNN-promptBackbone=Llama-2 70B, Shots=128-shots2024.0587.5
SCBackbone Model=Llama, Shot Count=16-shot2025.0587.42
SCModel=Llama, k-shot=162025.0587.42
char-CNN2026.0487.2
ICLBackbone=Llama-1 30B, Shots=128-shots2024.0587.1
FADS-ICLNumber of training shots (m)=64, LLM scale=1.5B2024.0587
kNN-promptingBackbone=GPT-2 0.8B, Shots=128-shots2024.0586.9
PCModel=Llama-2 7b, ICL Setting=1-shot2024.0586.81
kNN-promptingNumber of training shots (m)=64, LLM scale=1.5B2024.0586.6
FADS-ICLNumber of training shots (m)=16, LLM scale=1.5B2024.0586.5
DecTn (Shots)=162022.1286.4
kNN-promptBackbone=Llama-1 7B, Shots=128-shots2024.0586.4
SFTBackbone=RoBERTa_LARGE, Samples per class=162025.0686.36
SBERTcandidate pool size (m)=2562024.0586.3
ICLBackbone=Llama-1 7B, Shots=128-shots2024.0586.2
ICLBackbone=Llama-1 13B, Shots=128-shots2024.0586.2