初稿评审

毫不费力

读者会读完:多半是 · 行话密度:通篇大白话

发出去之前,对这份初稿的九项解读。

Jev 读到的 state

你好,我觉得你这个论文完全不行。虽然本科毕业论文我不该要求太高,但是你:1. 表格没有标题 2. 字体不一致 3. 参考文献时有时无,格式不统一 4. 第三章的实验和第四章看起来没一点关系 我给你的结果标注了,你下周三之前改完发给我

9
问题
1
次请求
203毫秒
往返耗时
1,035
输入 token
US$0.000043
费用
jev-1.13.0
模型

领衔的几条解读

清晰度

Score

第一次读的人知不知道它在说什么。

64%
3.57 满分 4 · 毫不费力

行话密度

Score

每段里的流行词数量。

89%
0.14 满分 4 · 通篇大白话

读者会读完

Noul

唯一重要的参与度指标。

多半是
答案为「是」的概率
83%
抛硬币

同一次请求里的其余部分

具体程度

Score

具体主张对抗抽象说辞。

72%
3.29 满分 4 · 大多很具体

阅读门槛

Score

这段文字假定读者受过多少教育。

85%
2.88 满分 4 · 本科

语域

Choice

读者听到的那把嗓音。

99%
口语化 占 100%,共 5 个选项
100% 口语化

最强的部分

Choice

哪里已经奏效了。

50%
结尾 占 58%,共 6 个选项
58% 结尾

含有废话

Noul

删掉也不损失什么的句子。

多半是
答案为「是」的概率
87%
抛硬币

提出了真主张

Noul

有没有断言任何可被证伪的东西。

几乎肯定是
答案为「是」的概率
96%
抛硬币

分享这条解读

拿到链接的人都能看,并且会出现在「浏览」里。

发到 X
产生这条结果的确切请求

POST https://api.typesafe.ai/v1/systemone

{
  "state": "你好,我觉得你这个论文完全不行。虽然本科毕业论文我不该要求太高,但是你:1. 表格没有标题 2. 字体不一致 3. 参考文献时有时无,格式不统一 4. 第三章的实验和第四章看起来没一点关系 我给你的结果标注了,你下周三之前改完发给我",
  "model": "jev-latest",
  "questions": {
    "clarity": {
      "type": "score",
      "instructions": "How clear is this writing to a first-time reader?",
      "criteria": [
        "Impenetrable",
        "Needs re-reading",
        "Followable with effort",
        "Clear",
        "Effortless"
      ]
    },
    "concreteness": {
      "type": "score",
      "instructions": "How concrete and specific is this writing?",
      "criteria": [
        "Pure abstraction, no specifics at all",
        "Mostly general claims",
        "A mix of specifics and generalities",
        "Mostly specific",
        "Every claim is grounded in a specific detail"
      ]
    },
    "jargon": {
      "type": "score",
      "instructions": "How heavy is the jargon and buzzword usage?",
      "criteria": [
        "Plain language throughout",
        "Occasional industry terms",
        "Noticeably jargon-heavy",
        "Dense with buzzwords",
        "Almost entirely buzzwords"
      ]
    },
    "reading_level": {
      "type": "score",
      "instructions": "What reading level does this writing assume?",
      "criteria": [
        "Primary school",
        "Middle school",
        "High school",
        "Undergraduate",
        "Specialist or academic"
      ]
    },
    "register": {
      "type": "choice",
      "instructions": "What voice is this written in?",
      "criteria": {
        "Conversational": "Sounds like a person talking to you",
        "Professional": "Polished and neutral",
        "Corporate": "Institutional, hedged, committee-written",
        "Academic": "Formal, citation-shaped, careful",
        "Promotional": "Selling, superlative-heavy"
      }
    },
    "strongest_element": {
      "type": "choice",
      "instructions": "What is the strongest thing about this draft?",
      "criteria": {
        "The opening": "It earns the next sentence",
        "The structure": "It is easy to navigate and follow",
        "The evidence": "Claims are supported by specifics",
        "The voice": "It sounds like a person with a point of view",
        "The ending": "It lands and tells you what to do",
        "Nothing": "No element stands out yet"
      }
    },
    "would_finish": {
      "type": "noul",
      "instructions": "Would an average reader finish this?",
      "criteria": {
        "true": "They would read to the end",
        "false": "They would stop partway"
      }
    },
    "has_filler": {
      "type": "noul",
      "instructions": "Does this contain sentences that could be deleted without losing meaning?"
    },
    "says_something": {
      "type": "noul",
      "instructions": "Does this make a specific claim a reader could disagree with?",
      "criteria": {
        "true": "It takes a position or states a checkable fact",
        "false": "It is unobjectionable and could have been written about anything"
      }
    }
  }
}

把 Jev 指向你自己的文本

九个类型化答案,一次请求,大约半秒。挑一个镜头,或者自己写问题。