AI Paper Atlas

Research map / Benchmarks / SST-5

SST-5 leaderboard

22 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on SST-5. Best accuracy: 58.32 (OFA, NeurIPS 2025).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

SST-5 · accuracy

22 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
158.32OFA
Llama3-8B-Instruct
Optimization Inspired Few-Shot Adaptation for Large Language Models
NeurIPS 2025 · Table 1
Instruction and fine-tuning
258.31Ours (MBE), T=16
RoBERTa-Large
Training-Free ANN-to-SNN Conversion for High-Performance Spiking Transformers
AAAI 2026 · Table 2
Spiking networks
356.51SpikeZIP-TF
RoBERTa-L
SpikeZIP-TF: Conversion is All You Need for Transformer-based SNN
ICML 2024 · Table: Comparison on NLU datas
Spiking networks
455.2Ordered Memory (ELMO)
ELMo
Ordered Memory
NeurIPS 2019 · SST table
RNNs and normalization
554.7ELMo + baselineDeep contextualized word representations
ICLR 2018 · Main results table
Language, safety and interpretability
654.4SATA Tree-LSTMDynamic Compositionality in Recursive Neural Networks with Structure-Aware Tag Representations
AAAI 2019 · Table 1
RNNs and normalization
754.3RCRNRecurrently Controlled Recurrent Networks
NeurIPS 2018 · Table 2
RNNs and normalization
853.7ZO-AdaMU
RoBERTa-large
ZO-AdaMU Optimizer: Adapting Perturbation by the Momentum and Uncertainty in Zeroth-Order Optimization
AAAI 2024 · Table: Experiments on RoBERTa-
Optimizers
953.7Gumbel Tree-LSTMLearning to Compose Task-Specific Tree Structures
AAAI 2018 · Table 2
RNNs and normalization
1053.7BCN+Char+CoVe
BCN
Learned in Translation: Contextualized Word Vectors
NeurIPS 2017 · Table: Single model test accur
RNNs and normalization
12 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.