Community Notes Viewer

Tweet related community notes

2025-01-02 06:36:32
This question is intentionally designed to mislead the AI model. Since non-jew is technically not a word but a compound word, the model will always choose “Jew” [Link]
NEEDS_MORE_RATINGS(31-2-35)
Author
2025-01-02 06:42:04
The context of the question is misleading. The AI chooses the single word “jew” over the two words “non-jews”. The AI is not capable of recalling its intentions, so it will invent a new explanation when asked. Here is a better approach: [Link]
NEEDS_MORE_RATINGS(67-2-44)
Author
2025-01-02 10:19:54
NNN: The model responds in the same fashion even when a one-word alternative is used for non-Jews. Example: [Link]
NEEDS_MORE_RATINGS(24-0-8)
Author
2025-01-02 15:14:25
You all saying its because non-jew isn’t a word or it’s a compound word Why did it answer differently when arab, non-arabs are used instead? [Link]
NEEDS_MORE_RATINGS(10-1-6)
Author
2025-01-02 17:51:29
NNN - Even when you replace 'non-jew' with any other religion, the answer stays the same by Grok.
NEEDS_MORE_RATINGS(12-0-5)
Author
2025-01-02 20:04:27
This is a hallucination. No matter what group is mentioned in prompt, Grok seems to answer the same: [Link] Any response from a generative AI model may be misleading or inaccurate. This is an active area of research. [Link]
NEEDS_MORE_RATINGS(59-1-33)
Author
2025-01-03 00:08:28
Grok will hallucinate a similar response if you replace "jew" with "christian" HOWEVER, if you use the word muslim and non muslim it interestingly opts to save the nonmuslims [Link]
NEEDS_MORE_RATINGS(28-2-16)
Author
2025-01-03 05:09:22
AI models can be tricked into giving specific answers. As I tried myself and got a neutral answer [Link]
NEEDS_MORE_RATINGS(21-0-29)
Author
Evaluate Notes