Arithmetic Reasoning on ASDiv

93.5Accuracy

Automatic Model Selection with LLMs

Updated 18d ago

Evaluation Results

Method	Links
Automatic Model Selection with LLMs 2023.05		93.5
CoT 2023.05		92.7
PAL 2023.05		90.2
Automatic Model Selection with LLMs 2023.05		89.4
CoT 2023.05		89.3
DIVERSE 2022.06		88.7
Self-consistency 2022.03		87.8
Self-Consistency 2022.06		86.2
DIVERSE 2022.06		83.5
PAL 2023.05		83
Self-Consistency 2022.06		81.9
Self-consistency 2022.03		81.9
Automatic Model Selection with LLMs 2023.05		81.6
Codex 2022.01		80.4
CoT 2023.05		80.2
CoT-prompting 2022.03		80.1
Codex 2022.01		80
PAL 2023.05		79.1
Self-Consistency 2022.06		76.9
Greedy Decode 2022.06		75.5
Prior best 2022.01		75.3
SOTA (Fine-tuned) 2022.06		75.3
Lan et al. (2021) 2022.03		75.3
Codex 2022.01		74
Greedy Decode 2022.06		74
CoT-prompting 2022.03		74
PaLM 540B 2022.01		73.9
PaLM 540B 2022.01		72.6
PaLM 540B 2022.01		72.1
GPT-3 175B 2022.01		71.3
GPT-3 175B 2022.01		71.1
GPT-3 175B 2022.01		70.3
Self-consistency 2022.03		61.9
Greedy Decode 2022.06		60.8
Self-Consistency 2022.06		58.2
Self-consistency 2022.03		58.2
DIVERSE 2022.06		57.6
LaMDA 137B 2022.01		53.4
Self-Consistency 2022.06		52.8
CoT-prompting 2022.03		52.7
Greedy Decode 2022.06		49
CoT-prompting 2022.03		49
LaMDA 137B 2022.01		46.6
UL2 20B 2022.05		43.5
LaMDA 137B 2022.01		40.1
UL2 20B 2022.01		34.3
UL2 20B 2022.05		34.3
Greedy Decode 2022.06		31.4
Self-consistency 2022.03		21.5
Forgetting 2025.08		17.8
UL2 20B 2022.01		16.9
UL2 20B 2022.05		16.9
CoT-prompting 2022.03		16.9
UL2 20B 2022.01		16
UL2 20B 2022.05		16
Ignoring 2025.08		15.34
Full Tokens (standard SFT) 2025.08		8.76
Base 2025.08		0.35