CSQA2
ReportedNo description yet: we only describe an eval set when we can do it correctly.
- Kind
- Published by a third party's leaderboard, collected by Epoch AI. The leaderboard is named and linked.
- Link
- https://arxiv.org/pdf/2201.05320
- Score shown
- Epoch's column "Score", as a percentage, higher is better. The source gives no error bars.
- Results
- 6 in all: 3 for open-weight models, 2 for closed ones, 1 where Epoch does not say. 1 linked to a model page here.
Best open-weight result: T5-11B, 67.8%, number 1 of 6 results.
| # | Model | Weights | Setting | Score | Date | Source |
|---|---|---|---|---|---|---|
| 1 | T5-11BGoogleno page match | open | default | 67.8% | not dated | CommonsenseQA 2.0: Exposing the Limits of AI through Gamification |
| 2 | T5-3BGoogleno page match | open | default | 60.2% | not dated | Decker: Double Check with Heterogeneous Knowledge for Commonsense Fact Verification |
| 3 | gpt-3.5-turbo-0613OpenAI | closed | Shots: few | 57.0% | not dated | arxiv.org |
| 4 | T5-Large | not stated | default | 54.6% | not dated | CommonsenseQA 2.0: Exposing the Limits of AI through Gamification |
| 5 | text-davinci-001OpenAI | closed | default | 52.9% | not dated | CommonsenseQA 2.0: Exposing the Limits of AI through Gamification |
| 6 | meta-llama/Llama-2-70b-hfEpoch: Llama-2-70b-hf, Meta AI | open | Shots: few | 50.0% | not dated | arxiv.org |
Open-weight rows are shaded. Dates are when the result was run or evaluated, where the data says. Data from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Epoch AI. Method