Multitask Language Understanding on BIG-bench-lite (24 tasks)
3,777ScorePaLM 540B
Evaluation Results
| Method | Links | |
|---|---|---|
| PaLM 540BParameters=540B, Protocol=1-shot2022.10 | 3,777 | |
| GLM-130BParameters=130B, Protocol=3-shot2022.10 | 1,512 | |
| GLM-130BParameters=130B, Protocol=1-shot2022.10 | 1,491 | |
| GLM-130BParameters=130B, Protocol=0-shot2022.10 | 1,331 | |
| GPT-3 175BParameters=175B, Protocol=3-shot2022.10 | 1,318 | |
| GPT-3 175BParameters=175B, Protocol=1-shot2022.10 | 1,134 | |
| PaLM 540BParameters=540B, Protocol=0-shot2022.10 | 805 | |
| GPT-3 13BParameters=13B, Protocol=3-shot2022.10 | 795 | |
| GPT-3 13BParameters=13B, Protocol=1-shot2022.10 | 543 | |
| GPT-3 6.7BParameters=6.7B, Protocol=3-shot2022.10 | 540 | |
| GPT-3 175BParameters=175B, Protocol=0-shot2022.10 | 435 | |
| GPT-3 6.7BParameters=6.7B, Protocol=1-shot2022.10 | 293 | |
| GPT-3 2.6BParameters=2.6B, Protocol=3-shot2022.10 | 183 | |
| GPT-3 13BParameters=13B, Protocol=0-shot2022.10 | 177 | |
| GPT-3 2.6BParameters=2.6B, Protocol=1-shot2022.10 | 71 | |
| GPT-3 2.6BParameters=2.6B, Protocol=0-shot2022.10 | 60 | |
| GPT-3 6.7BParameters=6.7B, Protocol=0-shot2022.10 | -6 |