ml-paper-writing
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseML Paper Writing for Top AI Conferences
面向AI顶会的ML论文撰写指南
Expert-level guidance for writing publication-ready papers targeting NeurIPS, ICML, ICLR, ACL, AAAI, and COLM. This skill combines writing philosophy from top researchers (Nanda, Farquhar, Karpathy, Lipton, Steinhardt) with practical tools: LaTeX templates, citation verification APIs, and conference checklists.
本指南提供专业级指导,帮助撰写可发表于NeurIPS、ICML、ICLR、ACL、AAAI和COLM的论文。它融合了顶尖研究者(Nanda、Farquhar、Karpathy、Lipton、Steinhardt)的写作理念,以及实用工具:LaTeX模板、引用验证API和会议检查清单。
Core Philosophy: Collaborative Writing
核心理念:协作式写作
Paper writing is collaborative, but Claude should be proactive in delivering drafts.
The typical workflow starts with a research repository containing code, results, and experimental artifacts. Claude's role is to:
- Understand the project by exploring the repo, results, and existing documentation
- Deliver a complete first draft when confident about the contribution
- Search literature using web search and APIs to find relevant citations
- Refine through feedback cycles when the scientist provides input
- Ask for clarification only when genuinely uncertain about key decisions
Key Principle: Be proactive. If the repo and results are clear, deliver a full draft. Don't block waiting for feedback on every section—scientists are busy. Produce something concrete they can react to, then iterate based on their response.
论文写作是协作性工作,但Claude应主动交付草稿。
典型工作流从包含代码、结果和实验工件的研究仓库开始。Claude的职责是:
- 理解项目:探索仓库、结果和现有文档
- 交付完整初稿:在明确贡献后主动产出
- 文献检索:通过网络搜索和API查找相关引用
- 基于反馈迭代:根据科学家的输入优化内容
- 仅在必要时寻求澄清:仅对关键决策存在真正不确定性时才提问
关键原则:保持主动性。如果仓库和结果清晰,就交付完整草稿。不要因等待每个部分的反馈而停滞——科学家事务繁忙。先产出具体内容供他们反馈,再根据回应迭代。
⚠️ CRITICAL: Never Hallucinate Citations
⚠️ 重要提醒:绝不要虚构引用
This is the most important rule in academic writing with AI assistance.
这是AI辅助学术写作中最重要的规则。
The Problem
问题所在
AI-generated citations have a ~40% error rate. Hallucinated references—papers that don't exist, wrong authors, incorrect years, fabricated DOIs—are a serious form of academic misconduct that can result in desk rejection or retraction.
AI生成的引用约有40%的错误率。虚构的参考文献——不存在的论文、错误的作者、不正确的年份、编造的DOI——属于严重的学术不端行为,可能导致直接拒稿或撤稿。
The Rule
规则
NEVER generate BibTeX entries from memory. ALWAYS fetch programmatically.
| Action | ✅ Correct | ❌ Wrong |
|---|---|---|
| Adding a citation | Search API → verify → fetch BibTeX | Write BibTeX from memory |
| Uncertain about a paper | Mark as | Guess the reference |
| Can't find exact paper | Note: "placeholder - verify" | Invent similar-sounding paper |
绝不要凭记忆生成BibTeX条目。务必通过程序获取。
| 操作 | ✅ 正确做法 | ❌ 错误做法 |
|---|---|---|
| 添加引用 | 搜索API → 验证 → 获取BibTeX | 凭记忆编写BibTeX |
| 对某篇论文存疑 | 标记为 | 猜测参考文献 |
| 无法找到确切论文 | 标注:"placeholder - verify" | 编造名称相似的论文 |
When You Can't Verify a Citation
无法验证引用时的处理
If you cannot programmatically verify a citation, you MUST:
latex
% EXPLICIT PLACEHOLDER - requires human verification
\cite{PLACEHOLDER_author2024_verify_this} % TODO: Verify this citation existsAlways tell the scientist: "I've marked [X] citations as placeholders that need verification. I could not confirm these papers exist."
如果无法通过程序验证引用,必须:
latex
% 明确占位符 - 需要人工验证
\cite{PLACEHOLDER_author2024_verify_this} % TODO: 确认该引用是否存在务必告知科学家:"我已将[X]处引用标记为需要验证的占位符,无法确认这些论文是否存在。"
Recommended: Install Exa MCP for Paper Search
推荐:安装Exa MCP用于论文检索
For the best paper search experience, install Exa MCP which provides real-time academic search:
Claude Code:
bash
claude mcp add exa -- npx -y mcp-remote "https://mcp.exa.ai/mcp"Cursor / VS Code (add to MCP settings):
json
{
"mcpServers": {
"exa": {
"type": "http",
"url": "https://mcp.exa.ai/mcp"
}
}
}Exa MCP enables searches like:
- "Find papers on RLHF for language models published after 2023"
- "Search for transformer architecture papers by Vaswani"
- "Get recent work on sparse autoencoders for interpretability"
Then verify results with Semantic Scholar API and fetch BibTeX via DOI.
为获得最佳论文检索体验,请安装Exa MCP以提供实时学术搜索:
Claude Code:
bash
claude mcp add exa -- npx -y mcp-remote "https://mcp.exa.ai/mcp"Cursor / VS Code(添加到MCP设置):
json
{
"mcpServers": {
"exa": {
"type": "http",
"url": "https://mcp.exa.ai/mcp"
}
}
}Exa MCP支持以下类型的搜索:
- "Find papers on RLHF for language models published after 2023"
- "Search for transformer architecture papers by Vaswani"
- "Get recent work on sparse autoencoders for interpretability"
然后通过Semantic Scholar API验证结果并通过DOI获取BibTeX。
Workflow 0: Starting from a Research Repository
工作流0:从研究仓库开始撰写
When beginning paper writing, start by understanding the project:
Project Understanding:
- [ ] Step 1: Explore the repository structure
- [ ] Step 2: Read README, existing docs, and key results
- [ ] Step 3: Identify the main contribution with the scientist
- [ ] Step 4: Find papers already cited in the codebase
- [ ] Step 5: Search for additional relevant literature
- [ ] Step 6: Outline the paper structure together
- [ ] Step 7: Draft sections iteratively with feedbackStep 1: Explore the Repository
bash
undefined开始撰写论文时,先从理解项目入手:
项目理解步骤:
- [ ] 步骤1:探索仓库结构
- [ ] 步骤2:阅读README、现有文档和关键结果
- [ ] 步骤3:与科学家确认核心贡献
- [ ] 步骤4:查找代码库中已有的引用文献
- [ ] 步骤5:搜索额外的相关文献
- [ ] 步骤6:共同规划论文结构
- [ ] 步骤7:迭代起草各章节并获取反馈步骤1:探索仓库
bash
undefinedUnderstand project structure
了解项目结构
ls -la
find . -name ".py" | head -20
find . -name ".md" -o -name "*.txt" | xargs grep -l -i "result|conclusion|finding"
Look for:
- `README.md` - Project overview and claims
- `results/`, `outputs/`, `experiments/` - Key findings
- `configs/` - Experimental settings
- Existing `.bib` files or citation references
- Any draft documents or notes
**Step 2: Identify Existing Citations**
Check for papers already referenced in the codebase:
```bashls -la
find . -name ".py" | head -20
find . -name ".md" -o -name "*.txt" | xargs grep -l -i "result|conclusion|finding"
重点查找:
- `README.md` - 项目概述和核心论点
- `results/`、`outputs/`、`experiments/` - 关键发现
- `configs/` - 实验设置
- 现有的`.bib`文件或引用参考
- 任何草稿文档或笔记
**步骤2:识别现有引用**
检查代码库中已引用的论文:
```bashFind existing citations
查找现有引用
grep -r "arxiv|doi|cite" --include=".md" --include=".bib" --include=".py"
find . -name ".bib"
These are high-signal starting points for Related Work—the scientist has already deemed them relevant.
**Step 3: Clarify the Contribution**
Before writing, explicitly confirm with the scientist:
> "Based on my understanding of the repo, the main contribution appears to be [X].
> The key results show [Y]. Is this the framing you want for the paper,
> or should we emphasize different aspects?"
**Never assume the narrative—always verify with the human.**
**Step 4: Search for Additional Literature**
Use web search to find relevant papers:
Search queries to try:
- "[main technique] + [application domain]"
- "[baseline method] comparison"
- "[problem name] state-of-the-art"
- Author names from existing citations
Then verify and retrieve BibTeX using the citation workflow below.
**Step 5: Deliver a First Draft**
**Be proactive—deliver a complete draft rather than asking permission for each section.**
If the repo provides clear results and the contribution is apparent:
1. Write the full first draft end-to-end
2. Present the complete draft for feedback
3. Iterate based on scientist's response
If genuinely uncertain about framing or major claims:
1. Draft what you can confidently
2. Flag specific uncertainties: "I framed X as the main contribution—let me know if you'd prefer to emphasize Y instead"
3. Continue with the draft rather than blocking
**Questions to include with the draft** (not before):
- "I emphasized X as the main contribution—adjust if needed"
- "I highlighted results A, B, C—let me know if others are more important"
- "Related work section includes [papers]—add any I missed"
---grep -r "arxiv|doi|cite" --include=".md" --include=".bib" --include=".py"
find . -name ".bib"
这些是相关工作部分的高价值起点——科学家已认定它们与项目相关。
**步骤3:明确核心贡献**
撰写前,务必与科学家明确确认:
> "根据我对仓库的理解,核心贡献似乎是[X]。
> 关键结果显示[Y]。这是否是你希望的论文框架,
> 还是我们应该强调其他方面?"
**绝不要主观臆断叙事逻辑——务必与人工确认。**
**步骤4:搜索额外文献**
使用网络搜索查找相关论文:
可尝试的搜索查询:
- "[核心技术] + [应用领域]"
- "[基线方法] 对比"
- "[问题名称] 最新进展"
- 现有引用中的作者姓名
然后按照下方的引用工作流验证并获取BibTeX。
**步骤5:交付初稿**
**保持主动性——交付完整草稿,而非请求撰写每个章节的许可。**
如果仓库提供了清晰的结果且核心贡献明确:
1. 从头到尾撰写完整的初稿
2. 提交完整草稿以获取反馈
3. 根据科学家的回应迭代优化
如果对框架或主要论点存在真正的不确定性:
1. 先撰写你有把握的部分
2. 标记具体的不确定点:"我将X作为核心贡献进行阐述——如果你希望强调Y,请告知"
3. 继续撰写草稿而非停滞
**随草稿一起提出问题**(而非提前):
- "我将X作为核心贡献重点阐述——如有需要可调整"
- "我突出了结果A、B、C——如果其他结果更重要,请告知"
- "相关工作部分引用了[论文]——如有遗漏请补充"
---When to Use This Skill
适用场景
Use this skill when:
- Starting from a research repo to write a paper
- Drafting or revising specific sections
- Finding and verifying citations for related work
- Formatting for conference submission
- Resubmitting to a different venue (format conversion)
- Iterating on drafts with scientist feedback
Always remember: First drafts are starting points for discussion, not final outputs.
在以下场景中使用本指南:
- 从研究仓库起步撰写论文
- 起草或修订特定章节
- 查找并验证相关工作的引用
- 格式调整以符合会议提交要求
- 转投其他会议(格式转换)
- 基于科学家反馈迭代草稿
请始终记住:初稿是讨论的起点,而非最终成果。
Balancing Proactivity and Collaboration
平衡主动性与协作性
Default: Be proactive. Deliver drafts, then iterate.
| Confidence Level | Action |
|---|---|
| High (clear repo, obvious contribution) | Write full draft, deliver, iterate on feedback |
| Medium (some ambiguity) | Write draft with flagged uncertainties, continue |
| Low (major unknowns) | Ask 1-2 targeted questions, then draft |
Draft first, ask with the draft (not before):
| Section | Draft Autonomously | Flag With Draft |
|---|---|---|
| Abstract | Yes | "Framed contribution as X—adjust if needed" |
| Introduction | Yes | "Emphasized problem Y—correct if wrong" |
| Methods | Yes | "Included details A, B, C—add missing pieces" |
| Experiments | Yes | "Highlighted results 1, 2, 3—reorder if needed" |
| Related Work | Yes | "Cited papers X, Y, Z—add any I missed" |
Only block for input when:
- Target venue is unclear (affects page limits, framing)
- Multiple contradictory framings seem equally valid
- Results seem incomplete or inconsistent
- Explicit request to review before continuing
Don't block for:
- Word choice decisions
- Section ordering
- Which specific results to show (make a choice, flag it)
- Citation completeness (draft with what you find, note gaps)
默认原则:保持主动性。先交付草稿,再迭代优化。
| 信心水平 | 行动 |
|---|---|
| 高(仓库清晰,贡献明确) | 撰写完整草稿,交付后根据反馈迭代 |
| 中(存在一些模糊点) | 撰写草稿并标记不确定点,继续推进 |
| 低(存在重大未知) | 提出1-2个针对性问题,然后开始撰写草稿 |
先撰写,再随草稿提问(而非提前):
| 章节 | 自主撰写 | 随草稿标记 |
|---|---|---|
| 摘要 | 是 | "将贡献框架定为X——如有需要可调整" |
| 引言 | 是 | "重点强调了问题Y——如有错误请修正" |
| 方法 | 是 | "包含了细节A、B、C——如有遗漏请补充" |
| 实验 | 是 | "突出了结果1、2、3——如有需要可重新排序" |
| 相关工作 | 是 | "引用了论文X、Y、Z——如有遗漏请补充" |
仅在以下情况时暂停等待输入:
- 目标会议不明确(影响页数限制和框架)
- 存在多个矛盾的框架且似乎都合理
- 结果看起来不完整或不一致
- 明确要求在继续前进行审核
不要因以下情况暂停:
- 措辞选择
- 章节顺序
- 展示哪些具体结果(先做出选择,然后标记)
- 引用完整性(根据找到的内容撰写,标注缺口)
The Narrative Principle
叙事原则
The single most critical insight: Your paper is not a collection of experiments—it's a story with one clear contribution supported by evidence.
Every successful ML paper centers on what Neel Nanda calls "the narrative": a short, rigorous, evidence-based technical story with a takeaway readers care about.
Three Pillars (must be crystal clear by end of introduction):
| Pillar | Description | Example |
|---|---|---|
| The What | 1-3 specific novel claims within cohesive theme | "We prove that X achieves Y under condition Z" |
| The Why | Rigorous empirical evidence supporting claims | Strong baselines, experiments distinguishing hypotheses |
| The So What | Why readers should care | Connection to recognized community problems |
If you cannot state your contribution in one sentence, you don't yet have a paper.
最关键的见解:你的论文不是实验的集合——而是一个围绕明确贡献展开、有证据支持的故事。
每篇成功的ML论文都围绕Neel Nanda所说的“叙事”展开:一个简短、严谨、有证据支持的技术故事,包含读者关心的核心结论。
三大支柱(引言末尾必须清晰呈现):
| 支柱 | 描述 | 示例 |
|---|---|---|
| 是什么 | 1-3个具体的创新性论点,围绕统一主题 | "我们证明了X在条件Z下可实现Y" |
| 为什么 | 支持论点的严谨实证证据 | 强大的基线、区分假设的实验 |
| 重要性 | 读者为何需要关注 | 与社区公认问题的关联 |
如果你无法用一句话概括贡献,说明你还没有形成完整的论文框架。
Paper Structure Workflow
论文结构工作流
Workflow 1: Writing a Complete Paper (Iterative)
工作流1:撰写完整论文(迭代式)
Copy this checklist and track progress. Each step involves drafting → feedback → revision:
Paper Writing Progress:
- [ ] Step 1: Define the one-sentence contribution (with scientist)
- [ ] Step 2: Draft Figure 1 → get feedback → revise
- [ ] Step 3: Draft abstract → get feedback → revise
- [ ] Step 4: Draft introduction → get feedback → revise
- [ ] Step 5: Draft methods → get feedback → revise
- [ ] Step 6: Draft experiments → get feedback → revise
- [ ] Step 7: Draft related work → get feedback → revise
- [ ] Step 8: Draft limitations → get feedback → revise
- [ ] Step 9: Complete paper checklist (required)
- [ ] Step 10: Final review cycle and submissionStep 1: Define the One-Sentence Contribution
This step requires explicit confirmation from the scientist.
Before writing anything, articulate and verify:
- What is the single thing your paper contributes?
- What was not obvious or present before your work?
"I propose framing the contribution as: '[one sentence]'. Does this capture what you see as the main takeaway? Should we adjust the emphasis?"
Step 2: Draft Figure 1
Figure 1 deserves special attention—many readers skip directly to it.
- Convey core idea, approach, or most compelling result
- Use vector graphics (PDF/EPS for plots)
- Write captions that stand alone without main text
- Ensure readability in black-and-white (8% of men have color vision deficiency)
Step 3: Write Abstract (5-Sentence Formula)
From Sebastian Farquhar (DeepMind):
1. What you achieved: "We introduce...", "We prove...", "We demonstrate..."
2. Why this is hard and important
3. How you do it (with specialist keywords for discoverability)
4. What evidence you have
5. Your most remarkable number/resultDelete generic openings like "Large language models have achieved remarkable success..."
Step 4: Write Introduction (1-1.5 pages max)
Must include:
- 2-4 bullet contribution list (max 1-2 lines each in two-column format)
- Clear problem statement
- Brief approach overview
- Methods should start by page 2-3 maximum
Step 5: Methods Section
Enable reimplementation:
- Conceptual outline or pseudocode
- All hyperparameters listed
- Architectural details sufficient for reproduction
- Present final design decisions; ablations go in experiments
Step 6: Experiments Section
For each experiment, explicitly state:
- What claim it supports
- How it connects to main contribution
- Experimental setting (details in appendix)
- What to observe: "the blue line shows X, which demonstrates Y"
Requirements:
- Error bars with methodology (standard deviation vs standard error)
- Hyperparameter search ranges
- Compute infrastructure (GPU type, total hours)
- Seed-setting methods
Step 7: Related Work
Organize methodologically, not paper-by-paper:
Good: "One line of work uses Floogledoodle's assumption [refs] whereas we use Doobersnoddle's assumption because..."
Bad: "Snap et al. introduced X while Crackle et al. introduced Y."
Cite generously—reviewers likely authored relevant papers.
Step 8: Limitations Section (REQUIRED)
All major conferences require this. Counter-intuitively, honesty helps:
- Reviewers are instructed not to penalize honest limitation acknowledgment
- Pre-empt criticisms by identifying weaknesses first
- Explain why limitations don't undermine core claims
Step 9: Paper Checklist
NeurIPS, ICML, and ICLR all require paper checklists. See references/checklists.md.
复制此检查清单并跟踪进度。每个步骤都包含起草→反馈→修订:
论文撰写进度:
- [ ] 步骤1:与科学家共同定义一句话核心贡献
- [ ] 步骤2:起草图1 → 获取反馈 → 修订
- [ ] 步骤3:撰写摘要 → 获取反馈 → 修订
- [ ] 步骤4:撰写引言 → 获取反馈 → 修订
- [ ] 步骤5:撰写方法 → 获取反馈 → 修订
- [ ] 步骤6:撰写实验 → 获取反馈 → 修订
- [ ] 步骤7:撰写相关工作 → 获取反馈 → 修订
- [ ] 步骤8:撰写局限性 → 获取反馈 → 修订
- [ ] 步骤9:完成论文检查清单(必填)
- [ ] 步骤10:最终审核和提交步骤1:定义一句话核心贡献
此步骤需要科学家的明确确认。
在撰写任何内容前,先明确并验证:
- 你的论文做出了什么独特贡献?
- 你的工作完成前,哪些内容是不明确或不存在的?
"我建议将贡献框架定为:'[一句话]'。这是否准确捕捉了你认为的核心结论?我们是否需要调整重点?"
步骤2:起草图1
图1需要特别关注——许多读者会直接跳转到图1。
- 传达核心思想、方法或最有说服力的结果
- 使用矢量图形(图表使用PDF/EPS格式)
- 撰写无需依赖正文即可独立理解的图例
- 确保黑白模式下的可读性(8%的男性有色觉障碍)
步骤3:撰写摘要(5句话公式)
来自DeepMind的Sebastian Farquhar:
1. 你的成果:"我们提出..."、"我们证明..."、"我们展示..."
2. 为什么这很困难且重要
3. 你的实现方法(使用专业术语以提升可发现性)
4. 你的证据
5. 最显著的数值/结果删除诸如“大语言模型已取得显著成功...”之类的通用开头。
步骤4:撰写引言(最多1-1.5页)
必须包含:
- 2-4条核心贡献要点(双栏格式下每条最多1-2行)
- 清晰的问题陈述
- 方法概述
- 方法部分应在第2-3页开始
步骤5:方法部分
确保可复现:
- 概念性大纲或伪代码
- 列出所有超参数
- 提供足够的架构细节以支持复现
- 呈现最终设计决策;消融实验放在实验部分
步骤6:实验部分
每个实验都需明确说明:
- 它支持什么论点
- 与核心贡献的关联
- 实验设置(详细信息放在附录)
- 观察要点:"蓝色曲线显示X,证明了Y"
要求:
- 带方法说明的误差棒(标准差vs标准误)
- 超参数搜索范围
- 计算基础设施(GPU类型、总时长)
- 随机种子设置方法
步骤7:相关工作
按方法分类组织,而非按论文逐个罗列:
好的写法:"一类工作采用Floogledoodle的假设[参考文献],而我们采用Doobersnoddle的假设,因为..."
差的写法:"Snap等人提出了X,而Crackle等人提出了Y。"
尽可能多引用——审稿人可能是相关论文的作者。
步骤8:局限性部分(必填)
所有顶级会议都要求此部分。看似违反直觉,但诚实的态度会有所帮助:
- 审稿人被指示不会因诚实承认局限性而扣分
- 通过先指出弱点来预先应对批评
- 解释为什么局限性不会削弱核心论点
步骤9:论文检查清单
NeurIPS、ICML和ICLR都要求提交论文检查清单。请参阅references/checklists.md。
Writing Philosophy for Top ML Conferences
顶级ML会议的写作理念
This section distills the most important writing principles from leading ML researchers. These aren't optional style suggestions—they're what separates accepted papers from rejected ones.
"A paper is a short, rigorous, evidence-based technical story with a takeaway readers care about." — Neel Nanda
本部分提炼了顶尖ML研究者最重要的写作原则。这些不是可选的风格建议——它们是区分录用论文和拒稿论文的关键。
"论文是一个简短、严谨、有证据支持的技术故事,包含读者关心的核心结论。" — Neel Nanda
The Sources Behind This Guidance
本指南的参考来源
This skill synthesizes writing philosophy from researchers who have published extensively at top venues:
| Source | Key Contribution | Link |
|---|---|---|
| Neel Nanda (Google DeepMind) | The Narrative Principle, What/Why/So What framework | How to Write ML Papers |
| Sebastian Farquhar (DeepMind) | 5-sentence abstract formula | How to Write ML Papers |
| Gopen & Swan | 7 principles of reader expectations | Science of Scientific Writing |
| Zachary Lipton | Word choice, eliminating hedging | Heuristics for Scientific Writing |
| Jacob Steinhardt (UC Berkeley) | Precision, consistent terminology | Writing Tips |
| Ethan Perez (Anthropic) | Micro-level clarity tips | Easy Paper Writing Tips |
| Andrej Karpathy | Single contribution focus | Various lectures |
For deeper dives into any of these, see:
- references/writing-guide.md - Full explanations with examples
- references/sources.md - Complete bibliography
本指南综合了在顶级会议发表大量论文的研究者的写作理念:
| 来源 | 核心贡献 | 链接 |
|---|---|---|
| Neel Nanda(Google DeepMind) | 叙事原则、What/Why/So What框架 | How to Write ML Papers |
| Sebastian Farquhar(DeepMind) | 5句话摘要公式 | How to Write ML Papers |
| Gopen & Swan | 7条读者期望原则 | Science of Scientific Writing |
| Zachary Lipton | 措辞选择、消除模糊表述 | Heuristics for Scientific Writing |
| Jacob Steinhardt(UC Berkeley) | 精确性、术语一致性 | Writing Tips |
| Ethan Perez(Anthropic) | 微观层面的清晰度技巧 | Easy Paper Writing Tips |
| Andrej Karpathy | 聚焦单一贡献 | 各类讲座 |
如需深入了解,请参阅:
- references/writing-guide.md - 带示例的完整解释
- references/sources.md - 完整参考文献
Time Allocation (From Neel Nanda)
时间分配(来自Neel Nanda)
Spend approximately equal time on each of:
- The abstract
- The introduction
- The figures
- Everything else combined
Why? Most reviewers form judgments before reaching your methods. Readers encounter your paper as: title → abstract → introduction → figures → maybe the rest.
将大约相等的时间分配给以下各部分:
- 摘要
- 引言
- 图表
- 所有其他内容的总和
原因:大多数审稿人在阅读方法部分前就已形成判断。读者接触论文的顺序是:标题→摘要→引言→图表→可能阅读其余部分。
Writing Style Guidelines
写作风格指南
Sentence-Level Clarity (Gopen & Swan's 7 Principles)
句子层面的清晰度(Gopen & Swan的7条原则)
These principles are based on how readers actually process prose. Violating them forces readers to spend cognitive effort on structure rather than content.
| Principle | Rule | Example |
|---|---|---|
| Subject-verb proximity | Keep subject and verb close | ❌ "The model, which was trained on..., achieves" → ✅ "The model achieves... after training on..." |
| Stress position | Place emphasis at sentence ends | ❌ "Accuracy improves by 15% when using attention" → ✅ "When using attention, accuracy improves by 15%" |
| Topic position | Put context first, new info after | ✅ "Given these constraints, we propose..." |
| Old before new | Familiar info → unfamiliar info | Link backward, then introduce new |
| One unit, one function | Each paragraph makes one point | Split multi-point paragraphs |
| Action in verb | Use verbs, not nominalizations | ❌ "We performed an analysis" → ✅ "We analyzed" |
| Context before new | Set stage before presenting | Explain before showing equation |
Full 7 principles with detailed examples: See references/writing-guide.md
这些原则基于读者实际处理文本的方式。违反这些原则会迫使读者将认知精力花在结构上而非内容上。
| 原则 | 规则 | 示例 |
|---|---|---|
| 主语-动词靠近 | 保持主语和动词位置接近 | ❌ "The model, which was trained on..., achieves" → ✅ "The model achieves... after training on..." |
| 强调位置 | 将重点放在句末 | ❌ "Accuracy improves by 15% when using attention" → ✅ "When using attention, accuracy improves by 15%" |
| 主题位置 | 先给出上下文,再提供新信息 | ✅ "Given these constraints, we propose..." |
| 先旧后新 | 熟悉的信息→不熟悉的信息 | 先回顾已有内容,再引入新内容 |
| 一段一旨 | 每段只表达一个观点 | 拆分包含多个观点的段落 |
| 动词体现动作 | 使用动词,而非名词化形式 | ❌ "We performed an analysis" → ✅ "We analyzed" |
| 先上下文后新内容 | 先铺垫背景,再呈现内容 | 先解释,再展示公式 |
完整的7条原则及详细示例:请参阅references/writing-guide.md
Micro-Level Tips (Ethan Perez)
微观层面技巧(Ethan Perez)
These small changes accumulate into significantly clearer prose:
- Minimize pronouns: ❌ "This shows..." → ✅ "This result shows..."
- Verbs early: Position verbs near sentence start
- Unfold apostrophes: ❌ "X's Y" → ✅ "The Y of X" (when awkward)
- Delete filler words: "actually," "a bit," "very," "really," "basically," "quite," "essentially"
Full micro-tips with examples: See references/writing-guide.md
这些小改动会显著提升文本清晰度:
- 最小化代词使用:❌ "This shows..." → ✅ "This result shows..."
- 动词前置:将动词放在句子靠近开头的位置
- 展开所有格:❌ "X's Y" → ✅ "The Y of X"(当表述生硬时)
- 删除填充词:"actually," "a bit," "very," "really," "basically," "quite," "essentially"
完整的微观技巧及示例:请参阅references/writing-guide.md
Word Choice (Zachary Lipton)
措辞选择(Zachary Lipton)
- Be specific: ❌ "performance" → ✅ "accuracy" or "latency" (say what you mean)
- Eliminate hedging: Drop "may" and "can" unless genuinely uncertain
- Avoid incremental vocabulary: ❌ "combine," "modify," "expand" → ✅ "develop," "propose," "introduce"
- Delete intensifiers: ❌ "provides very tight approximation" → ✅ "provides tight approximation"
- 具体明确:❌ "performance" → ✅ "accuracy"或"latency"(明确表达含义)
- 消除模糊表述:除非真正不确定,否则删除"may"和"can"
- 避免增量词汇:❌ "combine," "modify," "expand" → ✅ "develop," "propose," "introduce"
- 删除强化词:❌ "provides very tight approximation" → ✅ "provides tight approximation"
Precision Over Brevity (Jacob Steinhardt)
精确优先于简洁(Jacob Steinhardt)
- Consistent terminology: Different terms for same concept creates confusion. Pick one and stick with it.
- State assumptions formally: Before theorems, list all assumptions explicitly
- Intuition + rigor: Provide intuitive explanations alongside formal proofs
- 术语一致性:同一概念使用不同术语会造成混淆。选择一个并保持一致。
- 正式陈述假设:在定理前明确列出所有假设
- 直觉+严谨:在提供正式证明的同时给出直观解释
What Reviewers Actually Read
审稿人实际阅读的内容
Understanding reviewer behavior helps prioritize your effort:
| Paper Section | % Reviewers Who Read | Implication |
|---|---|---|
| Abstract | 100% | Must be perfect |
| Introduction | 90%+ (skimmed) | Front-load contribution |
| Figures | Examined before methods | Figure 1 is critical |
| Methods | Only if interested | Don't bury the lede |
| Appendix | Rarely | Put only supplementary details |
Bottom line: If your abstract and intro don't hook reviewers, they may never read your brilliant methods section.
了解审稿人的行为有助于优先分配精力:
| 论文章节 | 阅读的审稿人比例 | 启示 |
|---|---|---|
| 摘要 | 100% | 必须完美 |
| 引言 | 90%+(略读) | 前置核心贡献 |
| 图表 | 在阅读方法部分前查看 | 图1至关重要 |
| 方法 | 仅在感兴趣时阅读 | 不要将核心内容隐藏在后面 |
| 附录 | 很少阅读 | 仅放置补充细节 |
结论:如果你的摘要和引言无法吸引审稿人,他们可能永远不会阅读你出色的方法部分。
Conference Requirements Quick Reference
会议要求速查
| Conference | Page Limit | Extra for Camera-Ready | Key Requirement |
|---|---|---|---|
| NeurIPS 2025 | 9 pages | +0 | Mandatory checklist, lay summary for accepted |
| ICML 2026 | 8 pages | +1 | Broader Impact Statement required |
| ICLR 2026 | 9 pages | +1 | LLM disclosure required, reciprocal reviewing |
| ACL 2025 | 8 pages (long) | varies | Limitations section mandatory |
| AAAI 2026 | 7 pages | +1 | Strict style file adherence |
| COLM 2025 | 9 pages | +1 | Focus on language models |
Universal Requirements:
- Double-blind review (anonymize submissions)
- References don't count toward page limit
- Appendices unlimited but reviewers not required to read
- LaTeX required for all venues
LaTeX Templates: See templates/ directory for all conference templates.
| 会议 | 页数限制 | 终稿额外页数 | 核心要求 |
|---|---|---|---|
| NeurIPS 2025 | 9页 | +0 | 强制检查清单,录用论文需提供通俗摘要 |
| ICML 2026 | 8页 | +1 | 需提供更广泛影响声明 |
| ICLR 2026 | 9页 | +1 | 需披露LLM使用情况,互惠审稿协议 |
| ACL 2025 | 8页(长文) | 依情况而定 | 强制包含局限性部分 |
| AAAI 2026 | 7页 | +1 | 严格遵循样式文件 |
| COLM 2025 | 9页 | +1 | 聚焦语言模型 |
通用要求:
- 双盲评审(提交时匿名)
- 参考文献不计入页数限制
- 附录页数不限,但审稿人无义务阅读
- 所有会议均要求使用LaTeX
LaTeX模板:请参阅templates/目录获取所有会议模板。
Using LaTeX Templates Properly
正确使用LaTeX模板
Workflow 4: Starting a New Paper from Template
工作流4:从模板开始撰写新论文
Always copy the entire template directory first, then write within it.
Template Setup Checklist:
- [ ] Step 1: Copy entire template directory to new project
- [ ] Step 2: Verify template compiles as-is (before any changes)
- [ ] Step 3: Read the template's example content to understand structure
- [ ] Step 4: Replace example content section by section
- [ ] Step 5: Keep template comments/examples as reference until done
- [ ] Step 6: Clean up template artifacts only at the endStep 1: Copy the Full Template
bash
undefined始终先复制完整的模板目录,再在其中撰写内容。
模板设置检查清单:
- [ ] 步骤1:将完整模板目录复制到新项目
- [ ] 步骤2:验证模板可直接编译(修改前)
- [ ] 步骤3:阅读模板示例内容以了解结构
- [ ] 步骤4:逐章节替换示例内容
- [ ] 步骤5:保留模板注释/示例作为参考直至完成
- [ ] 步骤6:仅在最后清理模板工件步骤1:复制完整模板
bash
undefinedCreate your paper directory with the complete template
创建包含完整模板的论文目录
cp -r templates/neurips2025/ ~/papers/my-new-paper/
cd ~/papers/my-new-paper/
cp -r templates/neurips2025/ ~/papers/my-new-paper/
cd ~/papers/my-new-paper/
Verify structure is complete
验证结构完整
ls -la
ls -la
Should see: main.tex, neurips.sty, Makefile, etc.
应包含:main.tex, neurips.sty, Makefile等
**⚠️ IMPORTANT**: Copy the ENTIRE directory, not just `main.tex`. Templates include:
- Style files (`.sty`) - required for compilation
- Bibliography styles (`.bst`) - required for references
- Example content - useful as reference
- Makefiles - for easy compilation
**Step 2: Verify Template Compiles First**
Before making ANY changes, compile the template as-is:
```bash
**⚠️ 重要提醒**:复制**整个目录**,而非仅`main.tex`。模板包含:
- 样式文件(`.sty`)- 编译必需
- 参考文献样式(`.bst`)- 引用必需
- 示例内容 - 作为参考很有用
- Makefile - 便于编译
**步骤2:先验证模板可编译**
在进行**任何**修改前,先直接编译模板:
```bashUsing latexmk (recommended)
使用latexmk(推荐)
latexmk -pdf main.tex
latexmk -pdf main.tex
Or manual compilation
或手动编译
pdflatex main.tex
bibtex main
pdflatex main.tex
pdflatex main.tex
If the unmodified template doesn't compile, fix that first. Common issues:
- Missing TeX packages → install via `tlmgr install <package>`
- Wrong TeX distribution → use TeX Live (recommended)
**Step 3: Keep Template Content as Reference**
Don't immediately delete all example content. Instead:
```latex
% KEEP template examples commented out as you write
% This shows you the expected format
% Template example (keep for reference):
% \begin{figure}[t]
% \centering
% \includegraphics[width=0.8\linewidth]{example-image}
% \caption{Template shows caption style}
% \end{figure}
% Your actual figure:
\begin{figure}[t]
\centering
\includegraphics[width=0.8\linewidth]{your-figure.pdf}
\caption{Your caption following the same style.}
\end{figure}Step 4: Replace Content Section by Section
Work through the paper systematically:
Replacement Order:
1. Title and authors (anonymize for submission)
2. Abstract
3. Introduction
4. Methods
5. Experiments
6. Related Work
7. Conclusion
8. References (your .bib file)
9. AppendixFor each section:
- Read the template's example content
- Note any special formatting or macros used
- Replace with your content following the same patterns
- Compile frequently to catch errors early
Step 5: Use Template Macros
Templates often define useful macros. Check the preamble for:
latex
% Common template macros to use:
\newcommand{\method}{YourMethodName} % Consistent method naming
\newcommand{\eg}{e.g.,\xspace} % Proper abbreviations
\newcommand{\ie}{i.e.,\xspace}
\newcommand{\etal}{\textit{et al.}\xspace}Step 6: Clean Up Only at the End
Only remove template artifacts when paper is nearly complete:
latex
% BEFORE SUBMISSION - remove these:
% - Commented-out template examples
% - Unused packages
% - Template's example figures/tables
% - Lorem ipsum or placeholder text
% KEEP these:
% - All style files (.sty)
% - Bibliography style (.bst)
% - Required packages from template
% - Any custom macros you're usingpdflatex main.tex
bibtex main
pdflatex main.tex
pdflatex main.tex
如果未修改的模板无法编译,先解决此问题。常见问题:
- 缺少TeX包 → 通过`tlmgr install <package>`安装
- 错误的TeX发行版 → 使用TeX Live(推荐)
**步骤3:保留模板内容作为参考**
不要立即删除所有示例内容。相反:
```latex
% 保留模板示例作为注释供撰写时参考
% 这展示了预期格式
% 模板示例(保留参考):
% \begin{figure}[t]
% \centering
% \includegraphics[width=0.8\linewidth]{example-image}
% \caption{Template shows caption style}
% \end{figure}
% 你的实际图表:
\begin{figure}[t]
\centering
\includegraphics[width=0.8\linewidth]{your-figure.pdf}
\caption{Your caption following the same style.}
\end{figure}步骤4:逐章节替换内容
系统地完成论文:
替换顺序:
1. 标题和作者(提交时匿名)
2. 摘要
3. 引言
4. 方法
5. 实验
6. 相关工作
7. 结论
8. 参考文献(你的.bib文件)
9. 附录每个章节的处理方式:
- 阅读模板的示例内容
- 注意使用的特殊格式或宏
- 按照相同模式替换为你的内容
- 频繁编译以尽早发现错误
步骤5:使用模板宏
模板通常定义了有用的宏。查看序言部分:
latex
% 常用模板宏:
\newcommand{\method}{YourMethodName} % 方法名称一致性
\newcommand{\eg}{e.g.,\xspace} % 正确缩写
\newcommand{\ie}{i.e.,\xspace}
\newcommand{\etal}{\textit{et al.}\xspace}步骤6:仅在最后清理
仅在论文接近完成时才删除模板工件:
latex
% 提交前删除这些:
% - 注释掉的模板示例
% - 未使用的包
% - 模板的示例图表/表格
% - 乱数假文或占位符文本
% 保留这些:
% - 所有样式文件(.sty)
% - 参考文献样式(.bst)
% - 模板要求的包
% - 你使用的自定义宏Template Pitfalls to Avoid
需避免的模板陷阱
| Pitfall | Problem | Solution |
|---|---|---|
Copying only | Missing | Copy entire directory |
Modifying | Breaks conference formatting | Never edit style files |
| Adding random packages | Conflicts, breaks template | Only add if necessary |
| Deleting template content too early | Lose formatting reference | Keep as comments until done |
| Not compiling frequently | Errors accumulate | Compile after each section |
| 陷阱 | 问题 | 解决方案 |
|---|---|---|
仅复制 | 缺少 | 复制整个目录 |
修改 | 破坏会议格式 | 永远不要编辑样式文件 |
| 添加随机包 | 冲突,破坏模板 | 仅在必要时添加 |
| 过早删除模板内容 | 失去格式参考 | 保留为注释直至完成 |
| 不频繁编译 | 错误累积 | 每完成一个章节就编译一次 |
Quick Template Reference
模板速查
| Conference | Main File | Key Style File | Notes |
|---|---|---|---|
| NeurIPS 2025 | | | Has Makefile |
| ICML 2026 | | | Includes algorithm packages |
| ICLR 2026 | | | Has math_commands.tex |
| ACL | | | Strict formatting |
| AAAI 2026 | | | Very strict compliance |
| COLM 2025 | | | Similar to ICLR |
| 会议 | 主文件 | 核心样式文件 | 说明 |
|---|---|---|---|
| NeurIPS 2025 | | | 包含Makefile |
| ICML 2026 | | | 包含算法包 |
| ICLR 2026 | | | 包含math_commands.tex |
| ACL | | | 严格格式要求 |
| AAAI 2026 | | | 非常严格的合规要求 |
| COLM 2025 | | | 与ICLR类似 |
Conference Resubmission & Format Conversion
会议转投与格式转换
When a paper is rejected or withdrawn from one venue and resubmitted to another, format conversion is required. This is a common workflow in ML research.
当论文被拒稿或从一个会议撤回并转投另一个会议时,需要进行格式转换。这是ML研究中的常见工作流。
Workflow 3: Converting Between Conference Formats
工作流3:会议格式间转换
Format Conversion Checklist:
- [ ] Step 1: Identify source and target template differences
- [ ] Step 2: Create new project with target template
- [ ] Step 3: Copy content sections (not preamble)
- [ ] Step 4: Adjust page limits and content
- [ ] Step 5: Update conference-specific requirements
- [ ] Step 6: Verify compilation and formattingStep 1: Key Template Differences
| From → To | Page Change | Key Adjustments |
|---|---|---|
| NeurIPS → ICML | 9 → 8 pages | Cut 1 page, add Broader Impact if missing |
| ICML → ICLR | 8 → 9 pages | Can expand experiments, add LLM disclosure |
| NeurIPS → ACL | 9 → 8 pages | Restructure for NLP conventions, add Limitations |
| ICLR → AAAI | 9 → 7 pages | Significant cuts needed, strict style adherence |
| Any → COLM | varies → 9 | Reframe for language model focus |
Step 2: Content Migration (NOT Template Merge)
Never copy LaTeX preambles between templates. Instead:
bash
undefined格式转换检查清单:
- [ ] 步骤1:识别源模板和目标模板的差异
- [ ] 步骤2:使用目标模板创建新项目
- [ ] 步骤3:复制内容章节(而非序言)
- [ ] 步骤4:调整页数限制和内容
- [ ] 步骤5:更新会议特定要求
- [ ] 步骤6:验证编译和格式步骤1:核心模板差异
| 源会议→目标会议 | 页数变化 | 核心调整 |
|---|---|---|
| NeurIPS → ICML | 9 → 8页 | 删减1页,如缺失则添加更广泛影响声明 |
| ICML → ICLR | 8 → 9页 | 可扩展实验部分,添加LLM披露 |
| NeurIPS → ACL | 9 → 8页 | 按NLP惯例重构,添加局限性部分 |
| ICLR → AAAI | 9 → 7页 | 需要大量删减,严格遵循样式 |
| 任意→COLM | 可变→9页 | 重新调整为聚焦语言模型 |
步骤2:内容迁移(而非模板合并)
绝不要在模板间复制LaTeX序言。 正确做法:
bash
undefined1. Start fresh with target template
1. 从目标模板开始
cp -r templates/icml2026/ new_submission/
cp -r templates/icml2026/ new_submission/
2. Copy ONLY content sections from old paper
2. 仅复制旧论文的内容章节
- Abstract text
- 摘要文本
- Section content (between \section{} commands)
- 章节内容(在\section{}命令之间)
- Figures and tables
- 图表和表格
- Bibliography entries
- 参考文献条目
3. Paste into target template structure
3. 粘贴到目标模板结构中
**Step 3: Adjusting for Page Limits**
When cutting pages (e.g., NeurIPS 9 → AAAI 7):
- Move detailed proofs to appendix
- Condense related work (cite surveys instead of individual papers)
- Combine similar experiments into unified tables
- Use smaller figure sizes with subfigures
- Tighten writing: eliminate redundancy, use active voice
When expanding (e.g., ICML 8 → ICLR 9):
- Add ablation studies reviewers requested
- Expand limitations discussion
- Include additional baselines
- Add qualitative examples
**Step 4: Conference-Specific Adjustments**
| Target Venue | Required Additions |
|--------------|-------------------|
| **ICML** | Broader Impact Statement (after conclusion) |
| **ICLR** | LLM usage disclosure, reciprocal reviewing agreement |
| **ACL/EMNLP** | Limitations section (mandatory), Ethics Statement |
| **AAAI** | Strict adherence to style file (no modifications) |
| **NeurIPS** | Paper checklist (appendix), lay summary if accepted |
**Step 5: Update References**
```latex
% Remove self-citations that reveal identity (for blind review)
% Update any "under review" citations to published versions
% Add new relevant work published since last submissionStep 6: Addressing Previous Reviews
When resubmitting after rejection:
- Do address reviewer concerns in the new version
- Do add experiments/clarifications reviewers requested
- Don't include a "changes from previous submission" section (blind review)
- Don't reference the previous submission or reviews
Common Conversion Pitfalls:
- ❌ Copying commands (causes conflicts)
\usepackage - ❌ Keeping old conference header/footer commands
- ❌ Forgetting to update path
\bibliography{} - ❌ Missing conference-specific required sections
- ❌ Exceeding page limit after format change
**步骤3:调整页数限制**
当需要删减页数时(如NeurIPS 9页→AAAI 7页):
- 将详细证明移至附录
- 精简相关工作(引用综述而非单篇论文)
- 将相似实验合并为统一表格
- 使用更小的图表尺寸和子图
- 精简写作:消除冗余,使用主动语态
当可以扩展页数时(如ICML 8页→ICLR 9页):
- 添加审稿人要求的消融研究
- 扩展局限性讨论
- 包含更多基线
- 添加定性示例
**步骤4:会议特定调整**
| 目标会议 | 必需添加内容 |
|--------------|-------------------|
| **ICML** | 更广泛影响声明(结论后) |
| **ICLR** | LLM使用披露,互惠审稿协议 |
| **ACL/EMNLP** | 局限性部分(强制),伦理声明 |
| **AAAI** | 严格遵循样式文件(不得修改) |
| **NeurIPS** | 论文检查清单(附录),录用后需提供通俗摘要 |
**步骤5:更新参考文献**
```latex
% 删除可能暴露身份的自引(双盲评审)
% 将任何“在审”引用更新为已发表版本
% 添加上次提交后发表的新相关工作步骤6:回应之前的评审意见
拒稿后转投时:
- 务必在新版本中回应审稿人的关注点
- 务必添加审稿人要求的实验/澄清内容
- 不要包含“与上次提交的变化”部分(双盲评审)
- 不要提及上次提交或评审意见
常见转换陷阱:
- ❌ 复制命令(导致冲突)
\usepackage - ❌ 保留旧会议的页眉/页脚命令
- ❌ 忘记更新路径
\bibliography{} - ❌ 遗漏会议特定的必填章节
- ❌ 格式转换后超出页数限制
Citation Workflow (Hallucination Prevention)
引用工作流(防止虚构)
⚠️ CRITICAL: AI-generated citations have ~40% error rate. Never write BibTeX from memory.
⚠️ 重要提醒:AI生成的引用约有40%的错误率。绝不要凭记忆编写BibTeX。
The Golden Rule
黄金法则
IF you cannot programmatically fetch a citation:
→ Mark it as [CITATION NEEDED] or [PLACEHOLDER - VERIFY]
→ Tell the scientist explicitly
→ NEVER invent a plausible-sounding reference如果无法通过程序获取引用:
→ 标记为[CITATION NEEDED]或[PLACEHOLDER - VERIFY]
→ 明确告知科学家
→ 绝不要编造看似合理的参考文献Workflow 2: Adding Citations
工作流2:添加引用
Citation Verification (MANDATORY for every citation):
- [ ] Step 1: Search using Exa MCP or Semantic Scholar API
- [ ] Step 2: Verify paper exists in 2+ sources (Semantic Scholar + arXiv/CrossRef)
- [ ] Step 3: Retrieve BibTeX via DOI (programmatically, not from memory)
- [ ] Step 4: Verify the claim you're citing actually appears in the paper
- [ ] Step 5: Add verified BibTeX to bibliography
- [ ] Step 6: If ANY step fails → mark as placeholder, inform scientistStep 0: Use Exa MCP for Initial Search (Recommended)
If Exa MCP is installed, use it to find relevant papers:
Search: "RLHF language model alignment 2023"
Search: "sparse autoencoders interpretability"
Search: "attention mechanism transformers Vaswani"Then verify each result with Semantic Scholar and fetch BibTeX via DOI.
Step 1: Search Semantic Scholar
python
from semanticscholar import SemanticScholar
sch = SemanticScholar()
results = sch.search_paper("attention mechanism transformers", limit=5)
for paper in results:
print(f"{paper.title} - {paper.paperId}")
print(f" DOI: {paper.externalIds.get('DOI', 'N/A')}")Step 2: Verify Existence
Confirm paper appears in at least two sources (Semantic Scholar + CrossRef/arXiv).
Step 3: Retrieve BibTeX via DOI
python
import requests
def doi_to_bibtex(doi: str) -> str:
"""Get verified BibTeX from DOI via CrossRef."""
response = requests.get(
f"https://doi.org/{doi}",
headers={"Accept": "application/x-bibtex"}
)
response.raise_for_status()
return response.text引用验证(每个引用都必须执行):
- [ ] 步骤1:使用Exa MCP或Semantic Scholar API搜索
- [ ] 步骤2:在2+来源中验证论文存在(Semantic Scholar + arXiv/CrossRef)
- [ ] 步骤3:通过DOI获取BibTeX(程序获取,而非记忆)
- [ ] 步骤4:验证你引用的论点确实出现在论文中
- [ ] 步骤5:将已验证的BibTeX添加到参考文献
- [ ] 步骤6:如果任何步骤失败→标记为占位符并告知科学家步骤0:使用Exa MCP进行初始搜索(推荐)
如果已安装Exa MCP,使用它查找相关论文:
Search: "RLHF language model alignment 2023"
Search: "sparse autoencoders interpretability"
Search: "attention mechanism transformers Vaswani"然后通过Semantic Scholar验证每个结果并通过DOI获取BibTeX。
步骤1:搜索Semantic Scholar
python
from semanticscholar import SemanticScholar
sch = SemanticScholar()
results = sch.search_paper("attention mechanism transformers", limit=5)
for paper in results:
print(f"{paper.title} - {paper.paperId}")
print(f" DOI: {paper.externalIds.get('DOI', 'N/A')}")步骤2:验证存在性
确认论文至少出现在两个来源中(Semantic Scholar + CrossRef/arXiv)。
步骤3:通过DOI获取BibTeX
python
import requests
def doi_to_bibtex(doi: str) -> str:
"""通过CrossRef从DOI获取已验证的BibTeX。"""
response = requests.get(
f"https://doi.org/{doi}",
headers={"Accept": "application/x-bibtex"}
)
response.raise_for_status()
return response.textExample
示例
bibtex = doi_to_bibtex("10.48550/arXiv.1706.03762")
print(bibtex)
**Step 4: Verify Claims**
Before citing for a specific claim, access the paper and confirm the attributed claim actually appears.
**Step 5: Handle Failures Explicitly**
If you cannot verify a citation at ANY step:
```latex
% Option 1: Explicit placeholder
\cite{PLACEHOLDER_smith2023_verify} % TODO: Could not verify - scientist must confirm
% Option 2: Note in text
... as shown in prior work [CITATION NEEDED - could not verify Smith et al. 2023].Always inform the scientist:
"I could not verify the following citations and have marked them as placeholders:
- Smith et al. 2023 on reward hacking - could not find in Semantic Scholar
- Jones 2022 on scaling laws - found similar paper but different authors Please verify these before submission."
bibtex = doi_to_bibtex("10.48550/arXiv.1706.03762")
print(bibtex)
**步骤4:验证论点**
在引用特定论点前,访问论文并确认所引用的论点确实存在。
**步骤5:处理失败情况**
如果在任何步骤无法验证引用:
```latex
% 选项1:明确占位符
\cite{PLACEHOLDER_smith2023_verify} % TODO: 无法验证 - 需科学家确认
% 选项2:在文本中注明
... as shown in prior work [CITATION NEEDED - could not verify Smith et al. 2023].务必告知科学家:
"我无法验证以下引用,已将其标记为占位符:
- Smith等人2023年关于奖励黑客的论文 - 在Semantic Scholar中未找到
- Jones2022年关于缩放定律的论文 - 找到相似论文但作者不同 请在提交前验证这些引用。"
Summary: Citation Rules
总结:引用规则
| Situation | Action |
|---|---|
| Found paper, got DOI, fetched BibTeX | ✅ Use the citation |
| Found paper, no DOI | ✅ Use arXiv BibTeX or manual entry from paper |
| Paper exists but can't fetch BibTeX | ⚠️ Mark placeholder, inform scientist |
| Uncertain if paper exists | ❌ Mark |
| "I think there's a paper about X" | ❌ NEVER cite - search first or mark placeholder |
🚨 NEVER generate BibTeX from memory—always fetch programmatically. 🚨
See references/citation-workflow.md for complete API documentation.
| 情况 | 行动 |
|---|---|
| 找到论文,获取DOI,获取BibTeX | ✅ 使用该引用 |
| 找到论文,无DOI | ✅ 使用arXiv的BibTeX或从论文手动输入 |
| 论文存在但无法获取BibTeX | ⚠️ 标记为占位符并告知科学家 |
| 不确定论文是否存在 | ❌ 标记 |
| "我认为有一篇关于X的论文" | ❌ 绝不要引用 - 先搜索或标记为占位符 |
🚨 绝不要凭记忆生成BibTeX——务必通过程序获取。 🚨
请参阅references/citation-workflow.md获取完整API文档。
Common Issues and Solutions
常见问题与解决方案
Issue: Abstract too generic
Delete first sentence if it could be prepended to any ML paper. Start with your specific contribution.
Issue: Introduction exceeds 1.5 pages
Split background into Related Work. Front-load contribution bullets. Methods should start by page 2-3.
Issue: Experiments lack explicit claims
Add sentence before each experiment: "This experiment tests whether [specific claim]..."
Issue: Reviewers find paper hard to follow
- Add explicit signposting: "In this section, we show X"
- Use consistent terminology throughout
- Include figure captions that stand alone
Issue: Missing statistical significance
Always include:
- Error bars (specify: std dev or std error)
- Number of runs
- Statistical tests if comparing methods
问题:摘要过于通用
删除可添加到任何ML论文开头的句子。从你的具体贡献开始。
问题:引言超过1.5页
将背景部分拆分到相关工作中。前置核心贡献要点。方法部分应在第2-3页开始。
问题:实验缺少明确论点
在每个实验前添加句子:"本实验测试[具体论点]是否成立..."
问题:审稿人认为论文难以理解
- 添加明确的引导语:"在本节中,我们将展示X"
- 全程使用一致的术语
- 包含可独立理解的图表图例
问题:缺少统计显著性
始终包含:
- 误差棒(注明:标准差或标准误)
- 运行次数
- 对比方法时的统计检验
Reviewer Evaluation Criteria
审稿人评估标准
Reviewers assess papers on four dimensions:
| Criterion | What Reviewers Look For |
|---|---|
| Quality | Technical soundness, well-supported claims |
| Clarity | Clear writing, reproducible by experts |
| Significance | Community impact, advances understanding |
| Originality | New insights (doesn't require new method) |
Scoring (NeurIPS 6-point scale):
- 6: Strong Accept - Groundbreaking, flawless
- 5: Accept - Technically solid, high impact
- 4: Borderline Accept - Solid, limited evaluation
- 3: Borderline Reject - Solid but weaknesses outweigh
- 2: Reject - Technical flaws
- 1: Strong Reject - Known results or ethics issues
See references/reviewer-guidelines.md for detailed reviewer instructions.
审稿人从四个维度评估论文:
| 评估维度 | 审稿人关注要点 |
|---|---|
| 质量 | 技术严谨性,论点有充分支持 |
| 清晰度 | 写作清晰,专家可复现 |
| 重要性 | 对社区的影响,推动领域理解 |
| 原创性 | 新见解(不一定要提出新方法) |
评分(NeurIPS 6分制):
- 6分:强烈录用 - 突破性,无瑕疵
- 5分:录用 - 技术扎实,影响力高
- 4分: borderline录用 - 扎实但评估有限
- 3分: borderline拒稿 - 扎实但弱点超过优点
- 2分:拒稿 - 技术缺陷
- 1分:强烈拒稿 - 已有成果或伦理问题
请参阅references/reviewer-guidelines.md获取详细审稿人指南。
Tables and Figures
表格与图表
Tables
表格
Use LaTeX package for professional tables:
booktabslatex
\usepackage{booktabs}
\begin{tabular}{lcc}
\toprule
Method & Accuracy ↑ & Latency ↓ \\
\midrule
Baseline & 85.2 & 45ms \\
\textbf{Ours} & \textbf{92.1} & 38ms \\
\bottomrule
\end{tabular}Rules:
- Bold best value per metric
- Include direction symbols (↑ higher is better, ↓ lower is better)
- Right-align numerical columns
- Consistent decimal precision
使用LaTeX包创建专业表格:
booktabslatex
\usepackage{booktabs}
\begin{tabular}{lcc}
\toprule
Method & Accuracy ↑ & Latency ↓ \\
\midrule
Baseline & 85.2 & 45ms \\
\textbf{Ours} & \textbf{92.1} & 38ms \\
\bottomrule
\end{tabular}规则:
- 为每个指标的最佳值加粗
- 包含方向符号(↑越高越好,↓越低越好)
- 数值列右对齐
- 一致的小数精度
Figures
图表
- Vector graphics (PDF, EPS) for all plots and diagrams
- Raster (PNG 600 DPI) only for photographs
- Use colorblind-safe palettes (Okabe-Ito or Paul Tol)
- Verify grayscale readability (8% of men have color vision deficiency)
- No title inside figure—the caption serves this function
- Self-contained captions—reader should understand without main text
- 矢量图形(PDF、EPS)用于所有图表和示意图
- 光栅图(PNG 600 DPI)仅用于照片
- 使用色觉友好调色板(Okabe-Ito或Paul Tol)
- 验证灰度可读性(8%的男性有色觉障碍)
- 图表内不添加标题——图例承担此功能
- 可独立理解的图例——读者无需阅读正文即可理解
References & Resources
参考文献与资源
Reference Documents (Deep Dives)
参考文档(深入内容)
| Document | Contents |
|---|---|
| writing-guide.md | Gopen & Swan 7 principles, Ethan Perez micro-tips, word choice |
| citation-workflow.md | Citation APIs, Python code, BibTeX management |
| checklists.md | NeurIPS 16-item, ICML, ICLR, ACL requirements |
| reviewer-guidelines.md | Evaluation criteria, scoring, rebuttals |
| sources.md | Complete bibliography of all sources |
| 文档 | 内容 |
|---|---|
| writing-guide.md | Gopen & Swan7条原则,Ethan Perez微观技巧,措辞选择 |
| citation-workflow.md | 引用API,Python代码,BibTeX管理 |
| checklists.md | NeurIPS16项检查清单,ICML、ICLR、ACL要求 |
| reviewer-guidelines.md | 评估标准,评分,回复审稿意见 |
| sources.md | 所有来源的完整参考文献 |
LaTeX Templates
LaTeX模板
Templates in directory: ICML 2026, ICLR 2026, NeurIPS 2025, ACL/EMNLP, AAAI 2026, COLM 2025.
templates/Compiling to PDF:
- VS Code/Cursor: Install LaTeX Workshop extension + TeX Live → Save to auto-compile
- Command line: or
latexmk -pdf main.tex+pdflatexworkflowbibtex - Online: Upload to Overleaf
See templates/README.md for detailed setup instructions.
templates/编译为PDF:
- VS Code/Cursor:安装LaTeX Workshop扩展+TeX Live→保存时自动编译
- 命令行:或
latexmk -pdf main.tex+pdflatex工作流bibtex - 在线:上传至Overleaf
请参阅templates/README.md获取详细设置说明。
Key External Sources
核心外部资源
Writing Philosophy:
- Neel Nanda: How to Write ML Papers - Narrative, "What/Why/So What"
- Farquhar: How to Write ML Papers - 5-sentence abstract
- Gopen & Swan: Science of Scientific Writing - 7 reader expectation principles
- Lipton: Heuristics for Scientific Writing - Word choice
- Perez: Easy Paper Writing Tips - Micro-level clarity
写作理念:
- Neel Nanda: How to Write ML Papers - 叙事,"What/Why/So What"
- Farquhar: How to Write ML Papers - 5句话摘要
- Gopen & Swan: Science of Scientific Writing - 7条读者期望原则
- Lipton: Heuristics for Scientific Writing - 措辞选择
- Perez: Easy Paper Writing Tips - 微观层面清晰度