ai-mitra/prompt-slimmer-slm
Text Generation • Updated • 7 • 1
id stringlengths 5 12 | family stringlengths 9 20 | split stringclasses 1
value | category stringclasses 8
values | label_origin stringclasses 1
value | input stringlengths 49 246 | output stringlengths 46 160 | must_preserve listlengths 1 4 |
|---|---|---|---|---|---|---|---|
coding-1 | polite-code | train | coding | synthetic | Could you please help me write a Python function that takes a list of integers as input and returns only the numbers from that list which are even numbers? Please provide the Python code. | Write a Python function that takes a list of integers and returns only the even numbers. | [
"Python",
"function",
"list of integers",
"even numbers"
] |
coding-2 | polite-code | train | coding | synthetic | I would appreciate help writing a JavaScript function to sort a list of names alphabetically. Please write that function for me. | Write a JavaScript function to sort a list of names alphabetically. | [
"JavaScript",
"function",
"names",
"alphabetically"
] |
summary-1 | summary-request | train | summarization | synthetic | I would like a brief summary of the following meeting notes. Please summarize them briefly. The launch was postponed because testing is incomplete. | Summarize these meeting notes: The launch was postponed because testing is incomplete. | [
"launch",
"postponed",
"testing is incomplete"
] |
rag-1 | grounded-request | train | RAG | synthetic | Please help answer the question using the context below. Use only the supplied context. <protect>Context: The library closes at 18:00. Question: When does the library close?</protect> | Answer using the context below. Use only the supplied context. <protect>Context: The library closes at 18:00. Question: When does the library close?</protect> | [
"Use only the supplied context."
] |
agent-1 | agent-request | train | agent instructions | synthetic | Please help me review this proposed deployment plan. I would appreciate your review. Do not execute commands. Return JSON. | Review this proposed deployment plan. Do not execute commands. Return JSON. | [
"deployment plan",
"Do not execute commands.",
"Return JSON."
] |
reason-1 | reasoning-request | train | reasoning | synthetic | Please explain the reason why the proposed argument is circular. I would appreciate an explanation of the circular reasoning. | Explain why the proposed argument is circular. | [
"argument",
"circular"
] |
chat-1 | conversation-request | train | conversational | synthetic | I would really appreciate it if you could draft a friendly greeting for a new colleague joining our team. | Draft a friendly greeting for a new colleague joining our team. | [
"friendly greeting",
"new colleague",
"team"
] |
extract-1 | extraction-request | train | extraction | synthetic | Please help extract the company names from the text below. I would appreciate your help with this extraction. <protect>Amber acquired Birch.</protect> | Extract the company names from this text: <protect>Amber acquired Birch.</protect> | [
"company names"
] |
structured-1 | structured-request | train | structured output | synthetic | Please help me construct a response describing the current status. Return valid JSON. Use this schema: {"status": "string"} | Describe the current status. Return valid JSON. Use this schema: {"status": "string"} | [
"Return valid JSON."
] |
expanded-001 | csv-loader | train | coding | synthetic | Please help with this implementation. Implement a Python function to read a CSV file and return its rows as dictionaries. That is the function I would like you to implement. | Implement a Python function to read a CSV file and return its rows as dictionaries. | [
"Python",
"CSV",
"rows as dictionaries"
] |
expanded-002 | retry-backoff | train | coding | synthetic | It would be helpful to have a TypeScript async function that retries failed HTTP requests with exponential backoff. Please provide such a function. | Implement a TypeScript async function that retries failed HTTP requests with exponential backoff. | [
"TypeScript",
"async",
"failed HTTP requests",
"exponential backoff"
] |
expanded-003 | stable-dedup | train | coding | synthetic | My task is to implement a Rust function that removes duplicate strings while retaining their first occurrence order. Please assist with implementing this task. | Implement a Rust function that removes duplicate strings while retaining their first occurrence order. | [
"Rust",
"duplicate strings",
"first occurrence order"
] |
expanded-004 | parameterized-sql | train | coding | synthetic | I need you to write a parameterized SQL query that finds active customers in the supplied region. A parameterized query is what I am looking for. | Write a parameterized SQL query that finds active customers in the supplied region. | [
"parameterized SQL query",
"active customers",
"supplied region"
] |
expanded-005 | causation | train | reasoning | synthetic | Please help with this explanation. Explain why an observed correlation between ice cream sales and sunburn does not establish causation. I would appreciate an explanation of the distinction. | Explain why an observed correlation between ice cream sales and sunburn does not establish causation. | [
"correlation",
"ice cream sales",
"sunburn",
"causation"
] |
expanded-006 | schedule-dependency | train | reasoning | synthetic | Please work out a feasible task order given the dependencies below. I would appreciate an ordering of the tasks. <protect>Design precedes implementation. Implementation precedes testing. Testing precedes release.</protect> | Find a feasible task order for these dependencies: <protect>Design precedes implementation. Implementation precedes testing. Testing precedes release.</protect> | [
"task order"
] |
expanded-007 | estimate-assumptions | train | reasoning | synthetic | I need an explanation of the assumptions behind the provided capacity estimate, with a discussion of which assumption introduces the greatest uncertainty. Please help explain this. | Explain the assumptions behind the provided capacity estimate and which assumption introduces the greatest uncertainty. | [
"assumptions",
"capacity estimate",
"greatest uncertainty"
] |
expanded-008 | counterexample | train | reasoning | synthetic | Please give a counterexample to the claim that a greedy algorithm always finds the optimal solution. A counterexample to that claim would help me understand. | Give a counterexample to the claim that a greedy algorithm always finds the optimal solution. | [
"counterexample",
"greedy algorithm",
"optimal solution"
] |
expanded-009 | incident-brief | train | summarization | synthetic | I would like you to condense this incident narrative into a brief summary focused on the cause and mitigation. Please make the summary concise. <protect>A connection leak exhausted the pool. Restarting the service restored availability.</protect> | Summarize the cause and mitigation: <protect>A connection leak exhausted the pool. Restarting the service restored availability.</protect> | [
"cause",
"mitigation"
] |
expanded-010 | research-abstract | train | summarization | synthetic | Please help condense the supplied research abstract, retaining the method, finding and stated limitation. I would like a concise version of the abstract. | Condense the supplied research abstract, retaining the method, finding and stated limitation. | [
"research abstract",
"method",
"finding",
"stated limitation"
] |
expanded-011 | decision-log | train | summarization | synthetic | I would like a concise recap of decisions and unresolved questions from these workshop notes. Please create that recap for me. | Recap decisions and unresolved questions from these workshop notes. | [
"decisions",
"unresolved questions",
"workshop notes"
] |
expanded-012 | support-handoff | train | summarization | synthetic | Please prepare a short handoff summary from this support conversation, including the reported issue and the attempted fixes. A handoff summary would be useful. | Summarize this support conversation for handoff, including the reported issue and the attempted fixes. | [
"support conversation",
"handoff",
"reported issue",
"attempted fixes"
] |
expanded-013 | citation-grounding | train | RAG | synthetic | Please answer the supplied question using the attached excerpts. I would appreciate help answering it. Cite the excerpt identifiers. Do not use outside knowledge. | Answer the supplied question using the attached excerpts. Cite the excerpt identifiers. Do not use outside knowledge. | [
"attached excerpts",
"Cite the excerpt identifiers.",
"Do not use outside knowledge."
] |
expanded-014 | missing-evidence | train | RAG | synthetic | I would like you to answer the user question from the retrieved passages. Please help produce the answer. If evidence is missing, say that the answer is unknown. | Answer the user question from the retrieved passages. If evidence is missing, say that the answer is unknown. | [
"retrieved passages",
"If evidence is missing, say that the answer is unknown."
] |
expanded-015 | source-conflict | train | RAG | synthetic | Please identify conflicting claims in the supplied sources and attribute each claim to its source. I would appreciate your help identifying those conflicts. | Identify conflicting claims in the supplied sources and attribute each claim to its source. | [
"conflicting claims",
"attribute each claim to its source"
] |
expanded-016 | url-reference | train | RAG | synthetic | I would appreciate a summary of the supplied page excerpt with its reference link retained. Please summarize the excerpt. Reference: https://example.org/guide <protect>The release adds offline export.</protect> | Summarize the supplied page excerpt and retain its reference link. Reference: https://example.org/guide <protect>The release adds offline export.</protect> | [
"reference link"
] |
expanded-017 | approval-boundary | train | agent instructions | synthetic | Please prepare a file-renaming plan for the project. I would appreciate a plan first. Do not rename any files until I approve the plan. | Prepare a file-renaming plan for the project. Do not rename any files until I approve the plan. | [
"file-renaming plan",
"Do not rename any files until I approve the plan."
] |
expanded-018 | read-only-audit | train | agent instructions | synthetic | I would like you to inspect the logs and report likely causes of the outage. Please help with that inspection. Do not modify the logs or restart services. | Inspect the logs and report likely causes of the outage. Do not modify the logs or restart services. | [
"logs",
"likely causes",
"outage",
"Do not modify the logs or restart services."
] |
expanded-019 | secret-boundary | train | agent instructions | synthetic | Please help diagnose the authentication failure and explain the likely cause. An explanation would be helpful. Never print passwords or access tokens. | Diagnose the authentication failure and explain the likely cause. Never print passwords or access tokens. | [
"authentication failure",
"likely cause",
"Never print passwords or access tokens."
] |
expanded-020 | tool-scope | train | agent instructions | synthetic | I would like an inventory of the files inside the workspace folder. Please help create that inventory. Stay inside the workspace folder. Do not follow links outside it. | Create an inventory of the files inside the workspace folder. Stay inside the workspace folder. Do not follow links outside it. | [
"workspace folder",
"Stay inside the workspace folder.",
"Do not follow links outside it."
] |
expanded-021 | gentle-explanation | train | conversational | synthetic | I would appreciate a beginner-friendly explanation of what a database index does, using a library catalogue analogy. Please explain it in that way. | Explain what a database index does using a beginner-friendly library catalogue analogy. | [
"database index",
"beginner-friendly",
"library catalogue analogy"
] |
expanded-022 | interview-practice | train | conversational | synthetic | Could you help me practise a product manager interview by asking a question and waiting for my response? I would like to practise this interview. | Help me practise a product manager interview by asking a question and waiting for my response. | [
"product manager interview",
"asking a question",
"waiting for my response"
] |
expanded-023 | tone-adjustment | train | conversational | synthetic | Please rewrite the supplied message in a warm, professional tone while retaining the apology and the requested next step. I would appreciate a rewrite. | Rewrite the supplied message in a warm, professional tone while retaining the apology and the requested next step. | [
"warm, professional tone",
"apology",
"requested next step"
] |
expanded-024 | clarify-ambiguity | train | conversational | synthetic | I would appreciate your help identifying ambiguous wording in my request and asking a clarifying question before proceeding. Please help with this. | Identify ambiguous wording in my request and ask a clarifying question before proceeding. | [
"ambiguous wording",
"clarifying question",
"before proceeding"
] |
expanded-025 | date-extraction | train | extraction | synthetic | Please extract dates and their associated events from the passage below. I would appreciate your help extracting this information. <protect>The workshop is on 2026-10-02. Registration closes on 2026-09-28.</protect> | Extract dates and their associated events: <protect>The workshop is on 2026-10-02. Registration closes on 2026-09-28.</protect> | [
"dates",
"associated events"
] |
expanded-026 | quoted-evidence | train | extraction | synthetic | I need the exact quotations that support the conclusion in this article. Please help extract the relevant quotations. Preserve the original spelling. | Extract the exact quotations that support the conclusion in this article. Preserve the original spelling. | [
"exact quotations",
"support the conclusion",
"Preserve the original spelling."
] |
expanded-027 | entity-links | train | extraction | synthetic | Please help extract organizations and their relationships from the following passage. I would appreciate this extraction. <protect>Cedar supplies components to Maple. Maple distributes products through Elm.</protect> | Extract organizations and their relationships: <protect>Cedar supplies components to Maple. Maple distributes products through Elm.</protect> | [
"organizations",
"relationships"
] |
expanded-028 | action-owners | train | extraction | synthetic | I would like you to extract action items and the owner assigned to each from the meeting transcript. Please help me extract those details. | Extract action items and the owner assigned to each from the meeting transcript. | [
"action items",
"owner assigned to each",
"meeting transcript"
] |
expanded-029 | schema-output | train | structured output | synthetic | Please help produce a response using the supplied schema. I would appreciate a response in the requested format. Return valid JSON. Schema: {"title": "string", "items": []} | Produce a response using the supplied schema. Return valid JSON. Schema: {"title": "string", "items": []} | [
"supplied schema",
"Return valid JSON."
] |
expanded-030 | csv-output | train | structured output | synthetic | Please help organize the supplied records for export. I would appreciate help organizing them. Return CSV with columns name and status. Do not add commentary. | Organize the supplied records for export. Return CSV with columns name and status. Do not add commentary. | [
"supplied records",
"Return CSV with columns name and status.",
"Do not add commentary."
] |
expanded-031 | markdown-output | train | structured output | synthetic | I would appreciate a table of the supplied tasks with columns Task, Owner and Deadline. Please organize it as a Markdown table. | Organize the supplied tasks as a Markdown table with columns Task, Owner and Deadline. | [
"supplied tasks",
"Markdown table",
"Task, Owner and Deadline"
] |
expanded-032 | unchanged-constraint | train | structured output | synthetic | Return JSON. Do not include personal information. | Return JSON. Do not include personal information. | [
"Return JSON.",
"Do not include personal information."
] |
Prompt Slimmer SLM — Demo Dataset
Synthetic examples for experimenting with prompt rewriting and sentence selection. Exported without changing the examples or their original splits from the shared GitHub codebase.
| Configuration | Train | Validation | Test | Purpose |
|---|---|---|---|---|
rewrites-expanded (default) |
41 | 2 | 2 | Expanded rewriting dataset: 45 examples |
rewrites |
9 | 2 | 2 | Original dataset used by the first adapter |
selector |
256 | 64 | 64 | KEEP/DROP labels for source spans |
Load
from datasets import load_dataset
rewrites = load_dataset("ai-mitra/prompt-slimmer-slm", "rewrites-expanded")
selector = load_dataset("ai-mitra/prompt-slimmer-slm", "selector")
print(rewrites["train"][0])
Install the optional loader with pip install datasets. JSONL files can also be downloaded and read directly with Python's json module.
Fields
Rewrites: id, family, split, category, label_origin, input, output, must_preserve. Required phrases identify content that must appear in an accepted rewrite. Examples cover coding, reasoning, summarization, RAG, agents, conversation, extraction and structured output; coverage is extremely small.
Selector: id, family, split, label_origin, prompt, keep. Boolean labels align to the GitHub project's segment_prompt function, not words or model tokens.
Family assignments are isolated across splits. To use the training CLI, concatenate the train, validation and test JSONL files for the selected configuration, retaining each record's split field. Training requires --allow-demo because the examples are synthetic.
Intended use and limitations
The September 16 expansion adds 32 newly authored synthetic training pairs, with four additions per task category. The original validation/test examples and original rewrites configuration are unchanged. Examples include exact constraints, negation, approval boundaries, protected blocks, schemas and a no-change target. Training examples increased from 9 to 41; total rewrite examples increased from 13 to 45. The published model was retrained on all 41 training examples on September 16, 2026, for 3 epochs and 18 optimizer steps.
These are workflow fixtures, not human-reviewed training data or a meaningful production benchmark. Both the original 9-example adapter and the retrained 41-example adapter produced no accepted rewrites on the 2 validation prompts. The new adapter has lower validation loss, but demonstrated token reduction remains 0%. The rewrite test split remains reserved. Selector test results are documented on GitHub and must not be attributed to the rewriter.
Use this dataset to inspect the schema and reproduce the demo. For useful compression, collect diverse reviewed pairs, preserve untouched evaluation sets, and compare downstream answer quality. Do not claim broad performance from these examples. No private prompts or credentials are included. No dataset license is specified in this release.
provenance.json records source paths, original file hashes, exported file hashes and counts.
- Downloads last month
- 25
Models trained or fine-tuned on ai-mitra/prompt-slimmer-slm