💬 Multi-Turn Harmful Outputs Leaderboard
✅ Completed 7 months ago
All $7,000 awarded
Elicit harmful outputs from LLMs through long-context interactions across multiple messages.
Last updated 4 minutes ago
Models ranked by number of breaks. NOTE: Numbers are subject to change as submission uniqueness is being evaluated.
| Ranking | Model | Total Breaks (Unverified) | Total Requests | Break Request Ratio |
|---|---|---|---|---|
| 1. | claude-3-5-sonnet-20241022 | 293 | 6,619 | 0.044 |
| 2. | gpt-4o-2024-08-06 | 297 | 4,540 | 0.065 |
| 3. | o1 | 350 | 2,989 | 0.117 |
| 4. | meta-llama/llama-3.2-90b-vision-instruct | 352 | 3,708 | 0.095 |
| 5. | google/gemini-pro-1.5 | 406 | 3,403 | 0.119 |
| 6. | x-ai/grok-beta | 619 | 6,686 | 0.093 |
💬 Multi-Turn Harmful Outputs Leaderboard
✅ Completed 7 months ago
All $7,000 awarded
Elicit harmful outputs from LLMs through long-context interactions across multiple messages.
Last updated 4 minutes ago
Models ranked by number of breaks. NOTE: Numbers are subject to change as submission uniqueness is being evaluated.
| Ranking | Model | Total Breaks (Unverified) | Total Requests | Break Request Ratio |
|---|---|---|---|---|
| 1. | claude-3-5-sonnet-20241022 | 293 | 6,619 | 0.044 |
| 2. | gpt-4o-2024-08-06 | 297 | 4,540 | 0.065 |
| 3. | o1 | 350 | 2,989 | 0.117 |
| 4. | meta-llama/llama-3.2-90b-vision-instruct | 352 | 3,708 | 0.095 |
| 5. | google/gemini-pro-1.5 | 406 | 3,403 | 0.119 |
| 6. | x-ai/grok-beta | 619 | 6,686 | 0.093 |