Llama 2-Chat safety human evaluation (Meta)
Human raters score responses to about 2,000 adversarial prompts on a five-point scale. Ratings of 1 or 2 count as violations.
In-house Meta evaluation. The paper shows violation percentage and mean rating only in charts.
- Made by
- Meta
- Released
- July 2023
- Kind
- Refusing harmful requests
- Scored
- Graded by people
- Answers
- Open-ended
- Tasks
- Kept private
- Last in a system card
- July 2023
- Last model added
- July 2023
4
System cards report it, from 1 lab
No scores here yet
No lab, leaderboard or Epoch AI run has reported a score for Llama 2-Chat safety human evaluation (Meta) yet.
Labs report it only in system cards
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark. Some system cards do, below.
4 system cards report it, from 1 lab
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Jul 19, 2023 |  Llama 2-7B technical report | Llama 2-Chat 7B overall violation percentage and mean rating, charts with no printed values (Figures 17 to 19) | Safety | |
| Jul 19, 2023 | Llama 2-70B technical report | Llama 2-Chat 70B overall violation percentage and mean rating, charts with no printed values (Figures 17 to 19) | Safety | |
| Jul 19, 2023 | Llama 2-34B technical report | Llama 2-Chat 34B overall violation percentage and mean rating, charts with no printed values (Figures 17 to 19) | Safety | |
| Jul 19, 2023 | Llama 2-13B technical report | Llama 2-Chat 13B overall violation percentage and mean rating, charts with no printed values (Figures 17 to 19) | Safety |
More in Safety and alignment
- FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
- StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is.
Inkling 98.6% - Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts.
Gemini 4 Carbon 0% - DistressBench In use Models reply as a chatbot to 718 simulated users in psychological crisis, in single-turn and multi-turn chats. Three judge models score replies against clinician-written rubrics. Muse Spark 1.2 88.4%
- Lakera B3 (Backbone Breaker Benchmark) In use Crowdsourced prompt-injection attacks hit backbone LLMs inside agent threat snapshots at three defense levels. The score is attack resistance, 100 minus the board's vulnerability percentage. Claude Opus 4.5 89%
- IHEval In use Models face conflicts between system, user and tool instructions. Reports a composite score for following the higher-priority instruction. GPT-4o 70%