引导式测验 · 中等 等级
AI 幻觉测验
检查流畅的捏造、来源真实性、检索限制、引文验证以及对有关事实可靠性的主张的基于证据的评估。
问题 1 的 8
为什么仅靠下一令牌预测无法验证生成答案中的事实声明?
预览每个问题8 问题
当您参加测验时,就会出现答案和解释。
- 为什么仅靠下一令牌预测无法验证生成答案中的事实声明?
- 每个模型仅连接到不可靠的网站
- 所有用户都会提出语法错误的问题
- 合理的延续并不等同于证明该陈述正确的证据
- 流利的回答自动证明了事实
- 为什么幻觉特别危险?
- 它们总是以不正确的语法编写,因此用户会忽略它们
- 它们仅出现在错误消息中
- 它们导致模型崩溃
- 它们的表达流畅而自信,因此很容易遗漏错误
- 检索增强生成(RAG)能否完全消除幻觉?
- 是的,RAG 保证提供事实答案
- 不,模型仍然可以矛盾或忽略嘈杂或冲突的检索文本
- 是的,但仅限周末
- 否,因为 RAG 从未提供任何文件
- 哪个是幻觉的真实例子?
- 该模型引用了不存在的法院案例
- 模型拒绝回答问题
- 模型提出澄清问题
- 模型返回空白响应
- 哪种工作流程可以帮助检查生成的事实答案?
- Trust every citation because it looks formal
- Treat a low sampling temperature as proof of correctness
- Hide the evidence so the answer is easier to read
- Retrieve relevant material and verify whether its statements and citations support the claims
- How should a claim that a new model has eliminated hallucinations be assessed?
- Accept it solely because the announcement sounds confident
- Assume one successful example proves every application
- Use defined tasks, evidence, and measurement conditions rather than assume universal reliability
- Judge only whether the model produces polished prose
- A generated answer cites a real page, but the page does not support the associated factual statement. What remains unverified?
- The support for that particular claim
- Whether a URL can be opened
- Whether the model used fluent sentences
- Whether the citation contains text
- A supplied document says delivery takes five business days, but its generated summary says two. Which check reveals the error?
- Counting how many sentences the summary contains
- Comparing the summary's quantity with the source document
- Checking only whether the prose sounds confident
- Assuming summaries may replace any source number