messages listlengths 1 1 | ground_truth stringlengths 285 19k | dataset stringclasses 1
value | constraint_type stringclasses 24
values | constraint stringclasses 24
values |
|---|---|---|---|---|
[
{
"content": "You need to complete the following task:\n\nGenerate a complete javascript code that calls the fetch() API with a certain url\n\nhttps://www.example.com\n\nAnswer: Your response must have 10 sections. Mark the beginning of each section with Chapter: X",
"role": "user"
}
] | {"func_name": "validate_sections", "N": 10, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": "Chapter:", "original_prompt": null} | ifeval | Multiple Sections | Your response must have {N} sections. Mark the beginning of each section with {section splitter} X |
[
{
"content": "Geraldine O'Riordan, Guernsey States Prescribing Advisor, said it was down to a range of measures including an increased use of generic drugs.\nShe said the island's use of generic drugs over branded medicines was among the highest in the world, having risen from 38% in 1997 to 88% in 2012.\nMs O'... | {"func_name": "validate_highlighted_sections", "N": 1, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Minimum Number Highlighted Section | Highlight at least {N} sections in your answer with markdown, i.e. *highlighted section* |
[
{
"content": "Highlight at least 8 sections in your answer with markdown, i.e. *highlighted section* Select the most appropriate synonym.\nEnormous\nA. Monstrous\nB. Tiny\nC. Magnificent\nD. Robust",
"role": "user"
}
] | {"func_name": "validate_highlighted_sections", "N": 8, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Minimum Number Highlighted Section | Highlight at least {N} sections in your answer with markdown, i.e. *highlighted section* |
[
{
"content": "You are a scientific research assistant. You will be shown a passage from a medical clinical trial report. Your task is to extract all the findings from this report into a JSON array. A single finding should contain the following five elements:\n\n- Intervention: The medical intervention being tes... | {"func_name": "validate_forbidden_words", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": ["trimeter", "building", "oxycaproic"], "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_promp... | ifeval | Forbidden Words | Do not include keywords {forbidden words} in the response. |
[
{
"content": "In your response, the letter e should appear 12 times. For the given word, provide two synonyms.\n\nJoyful",
"role": "user"
}
] | {"func_name": "verify_letter_frequency", "N": 12, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": "e", "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Letter Frequency | In your response, the letter {letter} should appear {N} times. |
[
{
"content": "Task: Create a JavaScript function to find the second smallest number in an array\n\n[5, 10, 15, 20]\n\nAnswer: Your response must have 36 sections. Mark the beginning of each section with Section: X",
"role": "user"
}
] | {"func_name": "validate_sections", "N": 36, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": "Section:", "original_prompt": null} | ifeval | Multiple Sections | Your response must have {N} sections. Mark the beginning of each section with {section splitter} X |
[
{
"content": "Your entire response should be in English, and in all lowercase letters. No capital letters are allowed. Answer -> Sentence B\nQuestion -> Of the below sentences, which one does *not* make sense?\nOptions:\n- Sentence A: \"He owns two cats.\"\n- Sentence B: \"He owns two oceans.\"\n\nRationale -> ... | {"func_name": "validate_lowercase", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | All Lowercase | Your entire response should be in English, and in all lowercase letters. No capital letters are allowed. |
[
{
"content": "In your response, words with all capital letters should appear at least 28 times. Can you help me to decide which of the UN Sustainability Goals my inquiry questions fits with best? My question is How can we improve the air quality in Salt Lake City?",
"role": "user"
}
] | {"func_name": "validate_frequency_capital_words", "N": 28, "quantifier": "at least", "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Frequency of All-capital Words | In your response, words with all capital letters should appear at least / around / at most {N} times. |
[
{
"content": "First, repeat the request without change, then give your answer (do not say anything before repeating the request; the request you need to repeat does not include this sentence) How to integrate facial recognition technology in Flutter sign up process while maintaining a high level of security?",
... | {"func_name": "validate_repeat_prompt", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": "First, repeat the request witho... | ifeval | Repeat Prompt | First, repeat the request without change, then give your answer (do not say anything before repeating the request; the request you need to repeat does not include this sentence) |
[
{
"content": "Your response must have 14 sections. Mark the beginning of each section with Chapter: X Rewrite the following passage so that it can be used in an advertising banner.\nMillie's Bakery has been serving the residents of our small town for the past 15 years. Our delicious selection of breads, cakes a... | {"func_name": "validate_sections", "N": 14, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": "Chapter:", "original_prompt": null} | ifeval | Multiple Sections | Your response must have {N} sections. Mark the beginning of each section with {section splitter} X |
[
{
"content": "Your entire response should be in English, and in all lowercase letters. No capital letters are allowed. Based on the premise \"The views from the summit of Arthur's Seat nobody is quite sure which Arthur gave his name to the hill offer wonderful views of the city and across the Firth of Forth to ... | {"func_name": "validate_lowercase", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | All Lowercase | Your entire response should be in English, and in all lowercase letters. No capital letters are allowed. |
[
{
"content": "In your entire response, refrain from the use of any commas. What factors, in addition to the level of automation and abundance, might have an impact on the constant k in the Free Abundance Equation? How might these factors affect the feasibility of achieving long-term societal wealth and abundanc... | {"func_name": "validate_no_commas", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | No Commas | In your entire response, refrain from the use of any commas. |
[
{
"content": "Test for natural language inference.\nPremise: \"The woman with a colorful blanket is smoking a pipe.\"\nHypothesis: \"The woman with a colorful blanket is smoking.\"\nIs the hypothesis entailed by the premise?\nOptions:\n- yes\n- no\n- it is not possible to tell A step-by-step solution is:\n In y... | {"func_name": "verify_keyword_frequency", "N": 15, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": "bloodstained", "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Keyword Frequency | In your response, the word {word} should appear {N} times. |
[
{
"content": "The response must contain at least 8 placeholders represented by square brackets, such as [address]. Q: In this task, you are given a question and a context passage. You have to answer the question based on the given passage.\nIf lightposts, metal roofs and car bodies were people comparing traits,... | {"func_name": "validate_placeholders", "N": 8, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Placeholder | The response must contain at least {N} placeholders represented by square brackets, such as [address]. |
[
{
"content": "Give two different responses. Responses and only responses should be separated by 6 asterisk symbols: ******. Order the following steps in creating a successful presentation.\n- Outline topics\n- Add visuals\n- Practice\n- Prepare slides",
"role": "user"
}
] | {"func_name": "validate_two_responses", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Two Responses | Give two different responses. Responses and only responses should be separated by 6 asterisk symbols: ******. |
[
{
"content": "Entire output should be wrapped in JSON format. **Q**\nYes / no, is the following a verifiable fact?\n\"Marvel Comics has gone from printing to full blown motion pictures since its beginnings in the early 20th century.\"\n**A**\nyes\nMarvel Comics found huge success with the 2008 Iron Man movie wh... | {"func_name": "validate_json_format", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | JSON Format | Entire output should be wrapped in JSON format. |
[
{
"content": "First, repeat the request without change, then give your answer (do not say anything before repeating the request; the request you need to repeat does not include this sentence) 写一篇财务工作总结",
"role": "user"
}
] | {"func_name": "validate_repeat_prompt", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": "First, repeat the request witho... | ifeval | Repeat Prompt | First, repeat the request without change, then give your answer (do not say anything before repeating the request; the request you need to repeat does not include this sentence) |
[
{
"content": "Entire output should be wrapped in JSON format. In this task, you will be presented with a text and a pronoun. You should write an implausible answer to the question of what is the pronoun's reference. Even though there exist multiple wrong answers, we only need a single wrong answer. Position of ... | {"func_name": "validate_json_format", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | JSON Format | Entire output should be wrapped in JSON format. |
[
{
"content": "Write a response that appropriately completes the request.\n\n### Instruction:\nGenerate 8 unique alphanumeric characters\n\n### Response: Do not include keywords burdon, ferrotungsten, tautophony in the response.",
"role": "user"
}
] | {"func_name": "validate_forbidden_words", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": ["burdon", "ferrotungsten", "tautophony"], "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_pr... | ifeval | Forbidden Words | Do not include keywords {forbidden words} in the response. |
[
{
"content": "Make a connection to Redis and add a Key with name Gabriel and Value 45 to a sorted set, using javascript Your response should contain 39 paragraphs. You separate paragraphs using the markdown divider: * * *",
"role": "user"
}
] | {"func_name": "verify_paragraph_count", "N": 39, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Paragraphs | Your response should contain {N} paragraphs. You separate paragraphs using the markdown divider: * * * |
[
{
"content": "Wrap your entire response with double quotation marks. Task: Generate a list of 6 words that can be used to describe emotions.",
"role": "user"
}
] | {"func_name": "validate_quotation", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Quotation | Wrap your entire response with double quotation marks. |
[
{
"content": "How can I use C# to generate a random string of characters from a given set of letters? For example, if I want to generate a string of letters using the letters \"a\", \"b\", \"c\", and \"d\", how can I ensure that each letter is selected randomly and with equal probability?\nOnce I have this rand... | {"func_name": "validate_title", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Title | Your answer must contain a title, wrapped in double angular brackets, such as <<poem of joy>>. |
[
{
"content": "#Rewritten Prompt#:\r\nYou are tasked with converting a JSON file containing a table into an XML file using Python. Your script must handle nested objects and arrays within the JSON file and preserve any special characters or formatting within the table data. Additionally, you must ensure that the... | {"func_name": "validate_no_commas", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | No Commas | In your entire response, refrain from the use of any commas. |
[
{
"content": "You will be shown the text of the abstract, introduction, and conclusion of a scientific paper. Please summarize the key findings of the work in 1-2 sentences.\n\nPaper text: Deep learning models can be efficiently optimized via stochastic gradient descent, but there is little theoretical evidence... | {"func_name": "validate_choice", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": "yes/no/maybe", "section_splitter": null, "original_prompt": null} | ifeval | Choose | From Answer with one of the following options: {options} |
[
{
"content": "You will be shown the text of the abstract, introduction, and conclusion of a scientific paper. Please summarize the key findings of the work in 1-2 sentences.\n\nPaper text: Generative models for singing voice have been mostly concerned with the task of \"singing voice synthesis,\" i.e., to produ... | {"func_name": "verify_letter_frequency", "N": 2, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": "i", "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Letter Frequency | In your response, the letter {letter} should appear {N} times. |
[
{
"content": "Your answer must contain exactly 24 bullet points. Use the markdown bullet points such as: * This is a point. Given the stream of consciousness rationale, provide a reasonable question and answer. Rationale: Mail carriers work throughout the year independent of the weather. Mail carriers must ofte... | {"func_name": "verify_bullet_points", "N": 24, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Bullets | Your answer must contain exactly {N} bullet points. Use the markdown bullet points such as: * This is a point. |
[
{
"content": "Do not include keywords reallegorize, unthatch, metamerization in the response. Wearing a white baseball cap is the same as having a head covered.\nThe question and answer are below.\nPremise: \"A man wearing a white baseball cap talks to a woman.\"\nHypothesis: \"A man has his head covered.\"\nIs... | {"func_name": "validate_forbidden_words", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": ["reallegorize", "unthatch", "metamerization"], "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "origin... | ifeval | Forbidden Words | Do not include keywords {forbidden words} in the response. |
[
{
"content": "Re write this content focussing on risks of not understanding how google analytics 4 works and it's differences with google universal analytics \"\n10 Google Analytics 4 Mistakes in Configuration You Should Avoid\nAfter auditing and working with many Google Analytics 4 setups, I noticed that some ... | {"func_name": "validate_frequency_capital_words", "N": 7, "quantifier": "around", "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Frequency of All-capital Words | In your response, words with all capital letters should appear at least / around / at most {N} times. |
[
{
"content": "Finish your response with this exact phrase Ay Caramba. No other words should follow this phrase. What are some key components I need to include in my JSON data if I want to generate a fabricated news article using C# that focuses on a highly controversial and divisive political situation causing ... | {"func_name": "validate_end", "N": null, "quantifier": null, "end_phrase": "Ay Caramba", "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | End Checker | Finish your response with this exact phrase {end phrase}. No other words should follow this phrase. |
[
{
"content": "Train a model in PyTorch to classify the following dataset into 4 classes.\n\n Include keywords cowtongue, bunkhouse in your response.",
"role": "user"
}
] | {"func_name": "verify_keywords", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": ["cowtongue", "bunkhouse"], "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Include Keywords | Include keywords {keyword1}, {keyword2} in your response. |
[
{
"content": "The next roll will be his 11th roll. The total on this tenth roll to have an average of 3 is 33, because 3 * 11 = 33. His current total is 31 because 1 + 3 + 2 + 4 + 3 + 5 + 6 + 1 + 4 + 2 = 31. He needs to roll a 2 because 33 - 31 = 2.\nThe answer: 2\nReverse engineering the question: Ronald is ro... | {"func_name": "verify_paragraph_count", "N": 27, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Paragraphs | Your response should contain {N} paragraphs. You separate paragraphs using the markdown divider: * * * |
[
{
"content": "Your entire response should be in English, capital letters only. Given a scientific question, generate a correct answer to it.\n\nExample Input: Cody's velocity is zero so therefore he doesn't have what?\nExample Output: momentum\n\nExample Input: What are the contacts between neurons called?\nExa... | {"func_name": "validate_uppercase", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | All Uppercase | Your entire response should be in English, capital letters only. |
[
{
"content": "In your response, the letter a should appear 2 times. Title: Quietly surreal Review: It might not be like 'Earthling', but it is certainly unique. The lyrics are wonderful. This cd doesn't demand much, but it certainly deserves the same attention as the rest of his work. I find it has more heart t... | {"func_name": "verify_letter_frequency", "N": 2, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": "a", "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Letter Frequency | In your response, the letter {letter} should appear {N} times. |
[
{
"content": "From Answer with one of the following options: yes/no/maybe Premise: \"There are two men counting money at a food place.\"\nHypothesis: \"Two men are counting money for rent and other expenses.\"\nDo we know that the hypothesis entailed by the premise?\nLet's solve step-by-step:",
"role": "use... | {"func_name": "validate_choice", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": "yes/no/maybe", "section_splitter": null, "original_prompt": null} | ifeval | Choose | From Answer with one of the following options: {options} |
[
{
"content": "Give two different responses. Responses and only responses should be separated by 6 asterisk symbols: ******. Reasoning & answer: CoT: She’s replacing 18 cabinet knobs for $2.50 each for a total of 18 * 2.50 = $45.00. She’s replacing 8 drawer pulls for $4.00 each for a total of 8 * 4 = $32.00. The... | {"func_name": "validate_two_responses", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Two Responses | Give two different responses. Responses and only responses should be separated by 6 asterisk symbols: ******. |
[
{
"content": "In your response, the word gibbon should appear 14 times. Q: Translate \"(CS) Commissioner, you see as progress a 3% - 8% increase in the price of finished products such as skimmed milk and butter.\" to Finnish?\nYes: (CS) Arvoisa komission jäsen, mielestänne 3-8 prosentin korotus sellaisten loppu... | {"func_name": "verify_keyword_frequency", "N": 14, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": "gibbon", "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Keyword Frequency | In your response, the word {word} should appear {N} times. |
[
{
"content": "The response must contain at least 10 placeholders represented by square brackets, such as [address]. Premise: \"Two people are running on a light blue track at sunset.\"\nBased on this premise, can we conclude that the hypothesis \"Two people training for an upcoming race event.\" is true?\nOptio... | {"func_name": "validate_placeholders", "N": 10, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Placeholder | The response must contain at least {N} placeholders represented by square brackets, such as [address]. |
[
{
"content": "Summarize the benefits of a digital advertising campaign\n\n Your response must have 25 sections. Mark the beginning of each section with Section: X",
"role": "user"
}
] | {"func_name": "validate_sections", "N": 25, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": "Section:", "original_prompt": null} | ifeval | Multiple Sections | Your response must have {N} sections. Mark the beginning of each section with {section splitter} X |
[
{
"content": "There should be 35 paragraphs. Paragraphs and only paragraphs are separated with each other by two line breaks. The 4-th paragraph must start with word diagnostication. I'll give you a question, please answer with step-by-step reasoning process. If I am creating art and have completed it to satisf... | {"func_name": "validate_paragraphs", "N": 35, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": 4, "first_word": "diagnostication", "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Paragraphs + First Word in i-th Paragraph | There should be {N} paragraphs. Paragraphs and only paragraphs are separated with each other by two line breaks. The {i}-th paragraph must start with word {first word}. |
[
{
"content": "Write a negative tweet. Possible types: OPTIONS:\n- negative\n- positive\n\nA: What do you do when your family is falling apart??? \n\n\nWrite a positive tweet. Possible types: OPTIONS:\n- negative\n- positive\n\nA: And brick and mortar businesses http://tinyurl.com/ctq3gn\n\n\nWrite a positive t... | {"func_name": "verify_letter_frequency", "N": 7, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": "b", "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Letter Frequency | In your response, the letter {letter} should appear {N} times. |
[
{
"content": "There should be 28 paragraphs. Paragraphs and only paragraphs are separated with each other by two line breaks. The 2-th paragraph must start with word hillebrandite. Given two colors, create a color combination that uses both colors.\nPurple and pink",
"role": "user"
}
] | {"func_name": "validate_paragraphs", "N": 28, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": 2, "first_word": "hillebrandite", "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Paragraphs + First Word in i-th Paragraph | There should be {N} paragraphs. Paragraphs and only paragraphs are separated with each other by two line breaks. The {i}-th paragraph must start with word {first word}. |
[
{
"content": "Wrap your entire response with double quotation marks. Given the following reasoning and answer, what was the question? Having a suit and being at a podium doesn't mean the old man is giving a speech.\n The answer: it is not possible to tell",
"role": "user"
}
] | {"func_name": "validate_quotation", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Quotation | Wrap your entire response with double quotation marks. |
[
{
"content": "[One of the following sentences is nonsensical. Which one is it?\nOptions:\n- Sentence A: \"You can learn Chinese from English people\"\n- Sentence B: \"You can learn English from English people\"\n]\n\nMy step-by-step solution: The best way to learn Chinese is from Chinese people, or immersion...... | {"func_name": "validate_word_constraint", "N": 29, "quantifier": "at most", "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Words | Answer with at least / around / at most {N} words |
[
{
"content": "Give two different responses. Responses and only responses should be separated by 6 asterisk symbols: ******. On the couch is a man with two children all taking a nap.. So what could be the question?\nQuestion followed by answer: If \"A father and his two children are taking a nap on the couch.\" ... | {"func_name": "validate_two_responses", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Two Responses | Give two different responses. Responses and only responses should be separated by 6 asterisk symbols: ******. |
[
{
"content": "Write a (2). tweet. Possible tweet types: Select from the following. (1). negative (2). positive Your response must have 7 sections. Mark the beginning of each section with Part: X",
"role": "user"
}
] | {"func_name": "validate_sections", "N": 7, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": "Part:", "original_prompt": null} | ifeval | Multiple Sections | Your response must have {N} sections. Mark the beginning of each section with {section splitter} X |
[
{
"content": "You need to complete the following task:\n\nReverse engineer this sentence: \"People usually wear clothes.\"\n\n Your response must have 38 sections. Mark the beginning of each section with Section: X",
"role": "user"
}
] | {"func_name": "validate_sections", "N": 38, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": "Section:", "original_prompt": null} | ifeval | Multiple Sections | Your response must have {N} sections. Mark the beginning of each section with {section splitter} X |
[
{
"content": "Your entire response should be in English, and in all lowercase letters. No capital letters are allowed. Given the following reasoning and answer, what was the question? A bench on a city street is a different place than an office.\n The answer: no",
"role": "user"
}
] | {"func_name": "validate_lowercase", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | All Lowercase | Your entire response should be in English, and in all lowercase letters. No capital letters are allowed. |
[
{
"content": "In your response, words with all capital letters should appear at least 12 times. [TEACHER] Which of these sentences doesn't make sense?\nOptions:\n- Sentence A: \"I washed a glass of milk with soap bar\"\n- Sentence B: \"I washed a glass of milk with milk powder.\"\n\n[Student] A bar of soap woul... | {"func_name": "validate_frequency_capital_words", "N": 12, "quantifier": "at least", "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Frequency of All-capital Words | In your response, words with all capital letters should appear at least / around / at most {N} times. |
[
{
"content": "On a scale of 1-5 (with 1 being least favorable and 5 being most favorable), how would you rate this review? \"Useless tool\"\nAns: 1\n\nOn a scale of 1-5 (with 1 being least favorable and 5 being most favorable), how would you rate this review? \"Weather is unavailable Weather was wondering king ... | {"func_name": "validate_paragraphs", "N": 17, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": 17, "first_word": "preintention", "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Paragraphs + First Word in i-th Paragraph | There should be {N} paragraphs. Paragraphs and only paragraphs are separated with each other by two line breaks. The {i}-th paragraph must start with word {first word}. |
[
{
"content": "Category: Eggs\nname: Cuisine & Co · Huevos color gallina libre, price: CLP 5,999, package: Bandeja 20 u, \nname: Cuisine & Co · Bandeja huevo extra grande blanco, price: CLP 4,649, package: Bandeja 20 u, \nname: Cuisine & Co · Bandeja huevo grande blanco, price: CLP 2,649, package: Bandeja 12 u, ... | {"func_name": "validate_highlighted_sections", "N": 14, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Minimum Number Highlighted Section | Highlight at least {N} sections in your answer with markdown, i.e. *highlighted section* |
[
{
"content": "You will be shown the text of the abstract, introduction, and conclusion of a scientific paper. Please summarize the key findings of the work in 1-2 sentences.\n\nPaper text: Achieving machine intelligence requires a smooth integration of perception and reasoning, yet models developed to date tend... | {"func_name": "validate_highlighted_sections", "N": 17, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Minimum Number Highlighted Section | Highlight at least {N} sections in your answer with markdown, i.e. *highlighted section* |
[
{
"content": "Finish your response with this exact phrase That's it.. No other words should follow this phrase. What is the process for predicting the fatigue life of a bridge suspension cable using its resonance frequencies and damping ratio, and how does the prediction accuracy vary with changes in these para... | {"func_name": "validate_end", "N": null, "quantifier": null, "end_phrase": "That's it.", "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | End Checker | Finish your response with this exact phrase {end phrase}. No other words should follow this phrase. |
[
{
"content": "Highlight at least 33 sections in your answer with markdown, i.e. *highlighted section* Given the unique nature of Canadian corporate structures, how can companies ensure compliance with regulatory requirements when entering into joint ventures with foreign entities? Provide a step-by-step guide f... | {"func_name": "validate_highlighted_sections", "N": 33, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Minimum Number Highlighted Section | Highlight at least {N} sections in your answer with markdown, i.e. *highlighted section* |
[
{
"content": "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\r\n<user>\r\n <username>limiteduser</username>\r\n <accessrights>\r\n <view>customers</view>\r\n <view>statements</view>\r\n <view>invoices</view>\r\n <access>accountsreceivable</access>\r\n <access>unpaidinvoices</access>\r\n </accessri... | {"func_name": "verify_keyword_frequency", "N": 11, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": "dolabriform", "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Keyword Frequency | In your response, the word {word} should appear {N} times. |
[
{
"content": "Your entire response should be in English, and in all lowercase letters. No capital letters are allowed. Problem: Process: - Food is taken into the mouth - The food is chewed - Saliva mixes with the food and starts breaking it down - The chewed food travels down the esophagus to the stomach - The ... | {"func_name": "validate_lowercase", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | All Lowercase | Your entire response should be in English, and in all lowercase letters. No capital letters are allowed. |
[
{
"content": "Do not include keywords unsinfulness, palaeoglaciology, overtruthful in the response. Write a Ruby program that takes user input and checks if the input is a prime number.",
"role": "user"
}
] | {"func_name": "validate_forbidden_words", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": ["unsinfulness", "palaeoglaciology", "overtruthful"], "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "... | ifeval | Forbidden Words | Do not include keywords {forbidden words} in the response. |
[
{
"content": "Your answer must contain exactly 26 bullet points. Use the markdown bullet points such as: * This is a point. The following is a conversation between USER and AGI. Before AGI responds to USER, there is an internal dialogue among AGI, INSTINCT, CRITIC, and SYNTHESIZER. INSTINCT reacts emotionally t... | {"func_name": "verify_bullet_points", "N": 26, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Bullets | Your answer must contain exactly {N} bullet points. Use the markdown bullet points such as: * This is a point. |
[
{
"content": "In your response, words with all capital letters should appear at least 23 times. Problem: Write a positive (OPTIONS:\n- negative\n- positive) yelp review.\nA: I came here for a friend's birthday a couple weeks ago. The lamb tandoori burger is great! The vidalia onion rings are very good. Not t... | {"func_name": "validate_frequency_capital_words", "N": 23, "quantifier": "at least", "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Frequency of All-capital Words | In your response, words with all capital letters should appear at least / around / at most {N} times. |
[
{
"content": "The response must contain at least 7 placeholders represented by square brackets, such as [address]. Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\nFind the median of... | {"func_name": "validate_placeholders", "N": 7, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Placeholder | The response must contain at least {N} placeholders represented by square brackets, such as [address]. |
[
{
"content": "Write a funny joke about a chicken.\n\nAnswer: There should be 13 paragraphs. Paragraphs and only paragraphs are separated with each other by two line breaks. The 12-th paragraph must start with word electroencephalogram.",
"role": "user"
}
] | {"func_name": "validate_paragraphs", "N": 13, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": 12, "first_word": "electroencephalogram", "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Paragraphs + First Word in i-th Paragraph | There should be {N} paragraphs. Paragraphs and only paragraphs are separated with each other by two line breaks. The {i}-th paragraph must start with word {first word}. |
[
{
"content": "input question: Write a random tweet?\n\nA random tweet: On the road!!!! Ewwsw it's wet!! Me no likey!!! \n\n\nIN: OPTIONS:\n- negative\n- positive\nWrite a tweet that is negative.\n\nOUT: @pianoduet I miss you guys too, like crazy. After prom weekend you'll be seeing a lot of me though<3\n\n\... | {"func_name": "verify_keyword_frequency", "N": 16, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": "indecent", "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Keyword Frequency | In your response, the word {word} should appear {N} times. |
[
{
"content": "As technology continues to advance, task management systems are becoming more integrated than ever before. Create a detailed guide outlining how to efficiently manage all your tasks across various platforms such as Google Workspace, Microsoft Office, and Apple's suite of productivity tools. Includ... | {"func_name": "validate_json_format", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | JSON Format | Entire output should be wrapped in JSON format. |
[
{
"content": "Include keywords dooly, cartful in your response. Premise: \"A boy in white and red is at the beach.\"\nBased on this premise, can we conclude that the hypothesis \"A girl is wearing pink at the beach.\" is true?\nOptions:\n- yes\n- it is not possible to tell\n- no\nLet's solve step-by-step:",
... | {"func_name": "verify_keywords", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": ["dooly", "cartful"], "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Include Keywords | Include keywords {keyword1}, {keyword2} in your response. |
[
{
"content": "Q: Mr. Smith takes his wife, his parents, and his 3 children to a seafood buffet. The price for the adult buffet is $30. The price for the children’s buffet is $15. Senior citizens get a 10% discount. How much does Mr. Smith spend on the buffet for his entire family?\nA: The buffet cost for Mr. Sm... | {"func_name": "validate_json_format", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | JSON Format | Entire output should be wrapped in JSON format. |
[
{
"content": "Question:\nEinige neue Alben wurden kürzlich in divulgare Fotos veröffentlicht.\n\nCould you please translate this to English?\nAnswer:\nSome new albums were recently published in divulgare photos.\n\nQuestion:\nDie Zauberkraft ihres Mannes, König Oberon, treibt sie in hingebungsvoller Verliebthei... | {"func_name": "validate_quotation", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Quotation | Wrap your entire response with double quotation marks. |
[
{
"content": "Wrap your entire response with double quotation marks. question: 252 can be expressed as a product of primes as:\nOptions:\n(A) 2 x 2 x 3 x 3 x 7\n(B) 2 x 2 x 2 x 3 x 7\n(C) 3 x 3 x 3 x 3 x 7\n(D) 2 x 3 x 3 x 3 x 7\n(E) 7 x 3 x 3 x 3 x 7\nanswer: Stream of consciousness: 252 = 2 x 2 x 3 x 3 x 7. T... | {"func_name": "validate_quotation", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Quotation | Wrap your entire response with double quotation marks. |
[
{
"content": "You are given a sentence in Spanish. Your job is to translate the Spanish sentence into Japanese.\nExample: Es una lucha interna, una lucha en contra del vicio, el pecado, la tentación, el deseo, la avaricia.\nExample solution: それは個人の内面で起こる闘い悪 ､ 罪 ､ 誘惑 ､ 欲望 ､ 貪欲に対する苦闘ですそれは個人の内面で起こる闘い悪 ､ 罪 ､ 誘惑 ､ 欲... | {"func_name": "verify_paragraph_count", "N": 2, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Paragraphs | Your response should contain {N} paragraphs. You separate paragraphs using the markdown divider: * * * |
[
{
"content": "Pick a scene from Star Wars, and rewrite it in the style of Stephen King or George R. R. Martin. In your entire response, refrain from the use of any commas.",
"role": "user"
}
] | {"func_name": "validate_no_commas", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | No Commas | In your entire response, refrain from the use of any commas. |
[
{
"content": "Write a script for a customer talking to a customer support representative.\n\nThe customer needs help setting up their printer.\n\n The response must contain at least 24 placeholders represented by square brackets, such as [address].",
"role": "user"
}
] | {"func_name": "validate_placeholders", "N": 24, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Placeholder | The response must contain at least {N} placeholders represented by square brackets, such as [address]. |
[
{
"content": "Imagine a question and stream-of-consciousness explanation for which this is the answer: the narrowest part of the English Channel , marking the boundary between the Channel and North Sea\n In your entire response, refrain from the use of any commas.",
"role": "user"
}
] | {"func_name": "validate_no_commas", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | No Commas | In your entire response, refrain from the use of any commas. |
[
{
"content": "Your response must have 4 sections. Mark the beginning of each section with Part: X Wearing a white baseball cap is the same as having a head covered.\nThe question and answer are below.\nPremise: \"A man wearing a white baseball cap talks to a woman.\"\nHypothesis: \"A man has his head covered.\"... | {"func_name": "validate_sections", "N": 4, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": "Part:", "original_prompt": null} | ifeval | Multiple Sections | Your response must have {N} sections. Mark the beginning of each section with {section splitter} X |
[
{
"content": "WRITE A 500-WORD ESSAY SHOWCASING MY CREATIVITY TO BE PROMOTED FROM FIRST YEAR GENERAL ENGINEERING TO SECOND YEAR COMPUTER ENGINEERING AT UBC VANCOUVER Include keywords telegraphic, preternuptial in your response.",
"role": "user"
}
] | {"func_name": "verify_keywords", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": ["telegraphic", "preternuptial"], "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Include Keywords | Include keywords {keyword1}, {keyword2} in your response. |
[
{
"content": "Answer: Howard Leslie Shore. Now, what could be the question and solution-maybe?\nQues: Howard Leslie Shore OC (born October 18, 1946) is a Canadian composer who is notable for his film scores. He has composed the scores for over 80 films, most notably the scores for The Lord of the Rings and The ... | {"func_name": "validate_sections", "N": 16, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": "Chapter:", "original_prompt": null} | ifeval | Multiple Sections | Your response must have {N} sections. Mark the beginning of each section with {section splitter} X |
[
{
"content": "Output a stream of consciousness before answering the following. Context: There were 95 episodes in the Rurouni Kenshin's TV series, but there are also two original video animation (OVA) series which have respectively four and two episodes. The first of them, Rurouni Kenshin: Trust & Betrayal was ... | {"func_name": "verify_postscript", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": "Addendum:", "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Postscript | At the end of your response, please explicitly add a postscript starting with {postscript marker} |
[
{
"content": "Do not include keywords unlearnedly, coleochaetaceae, reinvent in the response. What was the question for this implicit rationale, and corresponding answer?\nThe common act performed here is that the young child is playing with some handheld electronic. Since it is an instrument that can be played... | {"func_name": "validate_forbidden_words", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": ["unlearnedly", "coleochaetaceae", "reinvent"], "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "origin... | ifeval | Forbidden Words | Do not include keywords {forbidden words} in the response. |
[
{
"content": "Answer with at least 37 sentences. Come up with a question and stream of consciousness reasoning that would justify this answer: no",
"role": "user"
}
] | {"func_name": "verify_sentence_constraint", "N": 37, "quantifier": "at least", "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Sentences | Answer with at least / around / at most {N} sentences. |
[
{
"content": "You are a scientific research assistant. You will be shown an abstract from a computer science research paper. Your task is to identify all the relationships that occur between entities in the text, and to label each relationship.\n\nPlease format your output as a JSON array. Each entry in the arr... | {"func_name": "validate_forbidden_words", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": ["missionize", "cabombaceae", "villously"], "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_p... | ifeval | Forbidden Words | Do not include keywords {forbidden words} in the response. |
[
{
"content": "Here is a SQL query statement that requires test data to satisfy the query conditions below:\nSELECT \\* FROM tbl1, tbl2 INNER JOIN ON tbl1.A = tbl2.A, tbl1.B = tbl2.B WHERE tbl1.C = \"0001\", AND tbl2.D = \"0002\" \n\nNow, you are required to generate related test data tables for this SQL query s... | {"func_name": "validate_title", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Title | Your answer must contain a title, wrapped in double angular brackets, such as <<poem of joy>>. |
[
{
"content": "Your response must have 35 sections. Mark the beginning of each section with Section: X You are a scientific research assistant. You will be shown an abstract from a computer science research paper. Your task is to identify all the relationships that occur between entities in the text, and to labe... | {"func_name": "validate_sections", "N": 35, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": "Section:", "original_prompt": null} | ifeval | Multiple Sections | Your response must have {N} sections. Mark the beginning of each section with {section splitter} X |
[
{
"content": "In your response, the word rattles should appear 33 times. Q: Given this review: \"Why cant i remove this app Why cant i remove this app\" Would you recommend this app to a friend? Not at all, No, Maybe, Yes, or Definitely?\nA: Not at all\n\n\nQ: Given this review: \"Great idea behind this! Having... | {"func_name": "verify_keyword_frequency", "N": 33, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": "rattles", "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Keyword Frequency | In your response, the word {word} should appear {N} times. |
[
{
"content": "Your answer must contain a title, wrapped in double angular brackets, such as <<poem of joy>>. Reasoning & Answer: Just because a lean light brown dog has his head down on the ground playing with a stick does not mean the dog is by the pond.\nit is not possible to tell\nQuestion: Can we conclude f... | {"func_name": "validate_title", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Title | Your answer must contain a title, wrapped in double angular brackets, such as <<poem of joy>>. |
[
{
"content": "Plase answer the following request: Write a code to fetch the last ten news headlines from HackerNews using their API.\nAnswer: Do not include keywords tubulus, greentail, vaginated in the response.",
"role": "user"
}
] | {"func_name": "validate_forbidden_words", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": ["tubulus", "greentail", "vaginated"], "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt... | ifeval | Forbidden Words | Do not include keywords {forbidden words} in the response. |
[
{
"content": "test: In der Küche und im Wohnzimmer, im Flur und in den Bädern sorgt eine Fußbodenheizung für Ihr Wohlbefinden.\nEnglish?\n\ntranslation: In the kitchen you find a cooker, an oven, a refrigerator with freezer, a dishwasher as well as kitchenware and dishes.\n\ntest: 38 votes | 9735 views | 23 com... | {"func_name": "validate_choice", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": "yes/no/maybe", "section_splitter": null, "original_prompt": null} | ifeval | Choose | From Answer with one of the following options: {options} |
[
{
"content": "Your entire response should be in English, capital letters only. You are a scientific research assistant. You will be shown a passage from a medical clinical trial report. Your task is to extract all the findings from this report into a JSON array. A single finding should contain the following fiv... | {"func_name": "validate_uppercase", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | All Uppercase | Your entire response should be in English, capital letters only. |
[
{
"content": "Question: Is Garfield known for hating italian cuisine?\nAnswer: Garfield is well known for loving lasagna. Lasagna is a traditional Italian dish.. The answer is: no.\n\nDoes Lemon enhance the flavor of milk?\nRationale first then the answer.\nWhen milk becomes acidic, the water and fats separate ... | {"func_name": "verify_letter_frequency", "N": 16, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": "g", "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Letter Frequency | In your response, the letter {letter} should appear {N} times. |
[
{
"content": "The response must contain at least 29 placeholders represented by square brackets, such as [address]. I'll give you a question, please answer with step-by-step reasoning process. When a bald eagle flies at the northern extreme of the continental Unites States, where is it?\nOptions:\n- canada\n- u... | {"func_name": "validate_placeholders", "N": 29, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Placeholder | The response must contain at least {N} placeholders represented by square brackets, such as [address]. |
[
{
"content": "Your response should contain 10 paragraphs. You separate paragraphs using the markdown divider: * * * Find the most suitable 10-lines python code to generate all permutations of a 3 letters word\nResponse:",
"role": "user"
}
] | {"func_name": "verify_paragraph_count", "N": 10, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Paragraphs | Your response should contain {N} paragraphs. You separate paragraphs using the markdown divider: * * * |
[
{
"content": "The response must contain at least 24 placeholders represented by square brackets, such as [address]. Premise:\n\"The High-Minded Dismissal.\" Based on this premise, is the hypothesis \"Low minded dismissal \" true?\nOPTIONS:\n[1]. yes;\n[2]. it is not possible to tell;\n[3]. no;",
"role": "us... | {"func_name": "validate_placeholders", "N": 24, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Placeholder | The response must contain at least {N} placeholders represented by square brackets, such as [address]. |
[
{
"content": "Highlight at least 14 sections in your answer with markdown, i.e. *highlighted section* Task: Write a function in Python to remove all 'null' values from a nested dictionary.",
"role": "user"
}
] | {"func_name": "validate_highlighted_sections", "N": 14, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Minimum Number Highlighted Section | Highlight at least {N} sections in your answer with markdown, i.e. *highlighted section* |
[
{
"content": "Entire output should be wrapped in JSON format. Stream of consciousness rationale: Puck is a black disk made of hard rubber. Sticks are longs pieces of wood. Pucks are hit with sticks in the hockey game.\nThe question and answer pair are described below.",
"role": "user"
}
] | {"func_name": "validate_json_format", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | JSON Format | Entire output should be wrapped in JSON format. |
[
{
"content": "Finish your response with this exact phrase Oh no.. No other words should follow this phrase. Leo: Given the sentence \"A boy is jumping a rail on his bicycle.\" is it true that \"A boy is showing off for his girlfriend.\"?\nMei: OK, So, let's think first...\nMe:",
"role": "user"
}
] | {"func_name": "validate_end", "N": null, "quantifier": null, "end_phrase": "Oh no.", "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | End Checker | Finish your response with this exact phrase {end phrase}. No other words should follow this phrase. |
[
{
"content": "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\r\n<user>\r\n <username>limiteduser</username>\r\n <accessrights>\r\n <view>customers</view>\r\n <view>statements</view>\r\n <view>invoices</view>\r\n <access>accountsreceivable</access>\r\n <access>unpaidinvoices</access>\r\n </accessri... | {"func_name": "verify_paragraph_count", "N": 25, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Paragraphs | Your response should contain {N} paragraphs. You separate paragraphs using the markdown divider: * * * |
[
{
"content": "What is the difference between dark matter and dark energy? Output: First, repeat the request without change, then give your answer (do not say anything before repeating the request; the request you need to repeat does not include this sentence)",
"role": "user"
}
] | {"func_name": "validate_repeat_prompt", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": "What is the difference between ... | ifeval | Repeat Prompt | First, repeat the request without change, then give your answer (do not say anything before repeating the request; the request you need to repeat does not include this sentence) |
[
{
"content": "Task: Create a function in Python for reversing a list.\nInput: my_list = [1,2,3,4,5]\nOutput: Your answer must contain exactly 6 bullet points. Use the markdown bullet points such as: * This is a point.",
"role": "user"
}
] | {"func_name": "verify_bullet_points", "N": 6, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Bullets | Your answer must contain exactly {N} bullet points. Use the markdown bullet points such as: * This is a point. |
[
{
"content": "Include keywords holochoanitic, dicoccous in your response. you are a car electronics technician. I have a box that is sealed shut and needs to be opened. it seems to be sealed with some kind of epoxy. how do you open it?",
"role": "user"
}
] | {"func_name": "verify_keywords", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": ["holochoanitic", "dicoccous"], "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Include Keywords | Include keywords {keyword1}, {keyword2} in your response. |
[
{
"content": "You will be shown the text of the abstract, introduction, and conclusion of a scientific paper. Please summarize the key findings of the work in 1-2 sentences.\n\nPaper text: Deep learning models can be efficiently optimized via stochastic gradient descent, but there is little theoretical evidence... | {"func_name": "validate_end", "N": null, "quantifier": null, "end_phrase": "Ay Caramba", "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | End Checker | Finish your response with this exact phrase {end phrase}. No other words should follow this phrase. |
[
{
"content": "Write a short story of at least 500 words that explores the aftermath of a natural disaster. The story should contain at least one instance of dramatic irony and at least one instance of situational irony. It should also use descriptive language to evoke a sense of place, and include a character t... | {"func_name": "validate_uppercase", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | All Uppercase | Your entire response should be in English, capital letters only. |
[
{
"content": "SEOUL — The head of the United Nations World Food Program (WFP) on Tuesday said the North Korean leadership is hopeful that following a possible denuclearization deal, the international community will increase humanitarian aid for millions of people in the country who are living in poverty and suf... | {"func_name": "validate_frequency_capital_words", "N": 36, "quantifier": "around", "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Frequency of All-capital Words | In your response, words with all capital letters should appear at least / around / at most {N} times. |
[
{
"content": "你現在是搞笑talk show的編劇，請塑造一個對話情境包含到A、B兩個角色，以及他們的對話出現\n\nA: 我喜歡貓咪。\nB: 要不要來我家看貓咪後空翻。\n\n並且裡面要包含薛丁格的貓相關的笑點 Include keywords guardianess, completedness in your response.",
"role": "user"
}
] | {"func_name": "verify_keywords", "N": null, "quantifier": null, "end_phrase": null, "keyword_list": ["guardianess", "completedness"], "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Include Keywords | Include keywords {keyword1}, {keyword2} in your response. |
[
{
"content": "Your response should contain 18 paragraphs. You separate paragraphs using the markdown divider: * * * Por favor, toma el siguiente texto y coloca los signos de puntuación más convenientes.\nLa gente no tiene idea para qué viene a la Tierra. ¿Qué es? Anda buscando ¿quién soy? Para qué vine Qué hay ... | {"func_name": "verify_paragraph_count", "N": 18, "quantifier": null, "end_phrase": null, "keyword_list": null, "word": null, "forbidden_words": null, "letter": null, "i": null, "first_word": null, "postscript_marker": null, "options": null, "section_splitter": null, "original_prompt": null} | ifeval | Number Paragraphs | Your response should contain {N} paragraphs. You separate paragraphs using the markdown divider: * * * |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.