ANLI
Harbor adapterAdversarial NLI (ANLI) is a dataset collected via an iterative, adversarial human-and-model-in-the-loop procedure.
Verified100%Knowledge & Reasoning
- Samples
- 169,265
- Released
- 2019
- Scores
- 12
Trivia QA
TriviaQA is a reading comprehension dataset containing over 650K question-answer-evidence triples.
How to run
ltb · trivia-qa
$ npm i -g lettokenburn-cli$ ltb bench pull trivia-qa --n 50$ ltb bench run ./trivia-qa -a claude-code -m anthropic/claude-opus-4-8$ ltb bench push
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
Scores
60 sourced entries| Harness | Source | Confidence | Link | |||
|---|---|---|---|---|---|---|
| Llama 2 70B | 87.6 | N/A | Self-reported | medium | N/A | arxiv.org |
| Claude 2 | 87.5 | N/A | Independent | medium | Jul 11, 2023 | www-files.anthropic.com |
| Claude 2 | 87.5 | N/A | Self-reported | medium | N/A | www-cdn.anthropic.com |
| GPT-4-0613 | 87.0 | N/A | Independent | medium | N/A | paperswithcode.com |
| Claude 1.3 | 86.7 | N/A | Independent | medium | Jul 11, 2023 | www-files.anthropic.com |
| Claude 1.3 | 86.7 | N/A | Self-reported | medium | N/A | www-cdn.anthropic.com |
| PaLM 2-L | 86.1 | N/A | Independent | medium | May 17, 2023 | arxiv.org |
| PaLM 2-L | 86.1 | N/A | Self-reported | medium | N/A | arxiv.org |
| LLaMA 65B | 86.0 | N/A | Self-reported | medium | N/A | arxiv.org |
| GPT-3.5-Turbo-1106 | 85.8 | N/A | Independent | medium | N/A | arxiv.org |
| Llama 2 70B | 85.0 | N/A | Independent | medium | Jul 18, 2023 | arxiv.org |
| GPT-4-0613 | 84.8 | N/A | Independent | medium | Mar 15, 2023 | arxiv.org |
| LLaMA 2 34B | 84.6 | N/A | Self-reported | medium | N/A | arxiv.org |
| LLaMA 33B | 83.8 | N/A | Self-reported | medium | N/A | arxiv.org |
| DeepSeek-V3 | 82.9 | N/A | Self-reported | medium | N/A | arxiv.org |
| Llama-3.1-405B-Instruct | 82.7 | N/A | Independent | medium | N/A | arxiv.org |
| Mixtral-8x7B-Instruct-v0.1 | 82.2 | N/A | Independent | medium | N/A | arxiv.org |
| PaLM 2-M | 81.7 | N/A | Independent | medium | May 17, 2023 | arxiv.org |
| PaLM 2-M | 81.7 | N/A | Self-reported | medium | N/A | arxiv.org |
| PaLM 540B | 81.4 | N/A | Independent | medium | Apr 5, 2022 | arxiv.org |
| PaLM 540B | 81.4 | N/A | Independent | medium | Apr 5, 2022 | arxiv.org |
| PaLM 540B | 81.4 | N/A | Self-reported | medium | N/A | jmlr.org |
| DeepSeek-V2 | 80.0 | N/A | Self-reported | medium | N/A | arxiv.org |
| Falcon 40B | 79.9 | N/A | Independent | medium | N/A | arxiv.org |
| LLaMA 2 13B | 79.6 | N/A | Self-reported | medium | N/A | arxiv.org |
| LLaMA 2 13B | 79.6 | N/A | Independent | medium | N/A | arxiv.org |
| Claude Instant 1.1 | 78.9 | N/A | Independent | medium | Jul 11, 2023 | www-files.anthropic.com |
| Claude Instant 1.1 | 78.9 | N/A | Self-reported | medium | N/A | www-cdn.anthropic.com |
| Claude Instant 1.2 | 78.7 | N/A | Self-reported | medium | N/A | anthropic.com |
| LLaMA 13B | 77.9 | N/A | Self-reported | medium | N/A | arxiv.org |
| PaLM 540B | 76.9 | N/A | Independent | medium | Apr 5, 2022 | arxiv.org |
| GLaM | 75.8 | N/A | Self-reported | medium | N/A | arxiv.org |
| PaLM 2-S | 75.2 | N/A | Independent | medium | May 17, 2023 | arxiv.org |
| Mistral-7B-v0.1 | 75.2 | N/A | Independent | medium | N/A | arxiv.org |
| PaLM 2-S | 75.2 | N/A | Self-reported | medium | N/A | arxiv.org |
| Phi-3-medium-128k-instruct | 73.9 | N/A | Self-reported | medium | N/A | arxiv.org |
| LLaMA 2 7B | 73.7 | N/A | Self-reported | medium | N/A | arxiv.org |
| MPT 30B | 73.6 | N/A | Independent | medium | N/A | arxiv.org |
| LLaMA 65B | 73.0 | N/A | Independent | medium | Feb 27, 2023 | arxiv.org |
| LLaMA 65B | 72.6 | N/A | Independent | medium | Feb 27, 2023 | arxiv.org |
| Gemma 7B | 72.3 | N/A | Independent | medium | N/A | arxiv.org |
| LLaMA 2 7B | 72.1 | N/A | Independent | medium | N/A | arxiv.org |
| Qwen2.5-72B-Instruct | 71.9 | N/A | Independent | medium | N/A | arxiv.org |
| LLaMA 65B | 71.6 | N/A | Independent | medium | Feb 27, 2023 | arxiv.org |
| GPT-3-175B | 71.2 | N/A | Independent | medium | May 28, 2020 | arxiv.org |
| text-davinci-001 | 71.2 | N/A | Independent | medium | N/A | arxiv.org |
| LLaMA 7B | 71.0 | N/A | Self-reported | medium | N/A | arxiv.org |
| Mistral 7B | 69.9 | N/A | Independent | medium | Oct 10, 2023 | arxiv.org |
| LLaMA 65B | 68.2 | N/A | Independent | medium | Feb 27, 2023 | arxiv.org |
| Meta-Llama-3-8B-Instruct | 67.7 | N/A | Independent | medium | N/A | arxiv.org |
Top 50 of 60 by score
Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.
Reproduced by the community
All runsLoading community runs…
About
TriviaQA is a reading comprehension dataset containing over 650K question-answer-evidence triples. TriviaQA includes 95K question-answer pairs authored by trivia enthusiasts and independently gathered evidence documents, six per question on average, that provide high quality distant supervision for answering the questions.
Available in EleutherAI lm-evaluation-harness as task family triviaqa (1 tasks).