CompassMTL 567M with Tailor

Reported on 4 benchmarks across 3 tasks · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing4 results

Sentence CompletiononHellaSwag
Accuracy· 2022-10-12
96.1
SOTA
Task Compass: Scaling Multi-task Pre-training with Task Prefix arXiv:2210.06277
Question AnsweringonSIQA
Accuracy· 2022-10-12
82.2
best: 83.2 (Unicorn 11B (fine-tuned))
Task Compass: Scaling Multi-task Pre-training with Task Prefix arXiv:2210.06277
Question AnsweringonPIQA
Accuracy· 2022-10-12
88.3
best: 90.1 (Unicorn 11B (fine-tuned))
Task Compass: Scaling Multi-task Pre-training with Task Prefix arXiv:2210.06277
Common Sense ReasoningonWinoGrande
Accuracy· 2022-10-12
90.5
best: 96.1 (ST-MoE-32B 269B (fine-tuned))
Task Compass: Scaling Multi-task Pre-training with Task Prefix arXiv:2210.06277