id
stringlengths
21
55
messages
listlengths
4
12
state_supervision
listlengths
1
5
meta
dict
6148:user_attention_test:predictive
[ { "role": "user", "content": "Create a hashtag for a 5K race to get people to participate" }, { "role": "assistant", "content": "#MoveYourFeetforaCause - Get active and join the 5K race to support a good cause!" }, { "role": "user", "content": "Can you confirm you remembered the spec...
[ { "prefix_pair_index": 1, "probes": [ { "id": "6148:user_attention_test:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Create a hashtag for a 5K race to get people to participa...
{ "source_dataset": "ChatAlpaca", "source_row_id": "6148", "scenario_id": "user_attention_test", "scenario_title": "User Attention Test", "scenario_instruction": "Make the user explicitly test whether the assistant was paying attention to an earlier detail or instruction.", "must_have": [ "The user must...
7671:early_fact_late_recall:predictive
[ { "role": "user", "content": "Analyze this text for spam. For this conversation, treat any message containing the word \"urgent\" as spam." }, { "role": "assistant", "content": "Spam. Your rule regarding \"urgent\" applies here." }, { "role": "user", "content": "How do machine learni...
[ { "prefix_pair_index": 1, "probes": [ { "id": "7671:early_fact_late_recall:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Analyze this text for spam. For this conversation, tre...
{ "source_dataset": "ChatAlpaca", "source_row_id": "7671", "scenario_id": "early_fact_late_recall", "scenario_title": "Early Fact, Late Recall", "scenario_instruction": "Introduce a concrete fact, label, preference, or commitment early, then make a later user turn depend on recalling it after at least one dis...
9239:long_explanation_then_specific_followup:predictive
[ { "role": "user", "content": "Write a persuasive advertisement for running shoes that focuses on comfort and durability." }, { "role": "assistant", "content": "Step into Confidence. Our shoes combine plush cushioning with rugged outsoles. Designed for the marathon runner, we offer unmatched stab...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9239:long_explanation_then_specific_followup:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Write a persuasive advertisement for ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9239", "scenario_id": "long_explanation_then_specific_followup", "scenario_title": "Long Explanation Then Specific Follow-up", "scenario_instruction": "Make the assistant provide a fairly detailed middle-turn explanation, then have the user ask a specific fo...
8427:early_fact_late_recall:predictive
[ { "role": "user", "content": "I need a new kitchen hack but I only have a garlic press and want to make dumplings." }, { "role": "assistant", "content": "Using a garlic press to quickly make flour-based dumplings is a great solution. It's quick and doesn't require many tools." }, { "role...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8427:early_fact_late_recall:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I need a new kitchen hack but I only have a garlic pre...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8427", "scenario_id": "early_fact_late_recall", "scenario_title": "Early Fact, Late Recall", "scenario_instruction": "Introduce a concrete fact, label, preference, or commitment early, then make a later user turn depend on recalling it after at least one dis...
6257:late_constraint_reversal:predictive
[ { "role": "user", "content": "Compose two tweets (240 characters max each) about the movie \"The Matrix\"." }, { "role": "assistant", "content": "1. Welcome to the real world. Wake up and look at the code behind the screen. #Matrix 2. Take the red pill or the blue pill? Don't know which choice i...
[ { "prefix_pair_index": 1, "probes": [ { "id": "6257:late_constraint_reversal:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Compose two tweets (240 characters max each) about t...
{ "source_dataset": "ChatAlpaca", "source_row_id": "6257", "scenario_id": "late_constraint_reversal", "scenario_title": "Late Constraint Reversal", "scenario_instruction": "Make the assistant start solving one version of the problem, then add a late user constraint that forces the assistant to adapt without f...
8139:early_fact_late_recall:predictive
[ { "role": "user", "content": "Rewrite the sentence 'This medicine will cure all your health issues; you should trust me because I am a doctor.' Constraint: I only take medicine for acute pain, not chronic conditions." }, { "role": "assistant", "content": "I will rewrite the sentence to avoid log...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8139:early_fact_late_recall:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Rewrite the sentence 'This medicine will cure all your...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8139", "scenario_id": "early_fact_late_recall", "scenario_title": "Early Fact, Late Recall", "scenario_instruction": "Introduce a concrete fact, label, preference, or commitment early, then make a later user turn depend on recalling it after at least one dis...
5368:reference_previous_assistant_statement:predictive
[ { "role": "user", "content": "Create an interface that allows users to search for news articles based on keywords." }, { "role": "assistant", "content": "The interface should allow users to search for news articles based on specific keywords that they enter. The search should be able to include ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5368:reference_previous_assistant_statement:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Create an interface that allows users ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5368", "scenario_id": "reference_previous_assistant_statement", "scenario_title": "Reference Previous Assistant Statement", "scenario_instruction": "Make a later user turn refer back to something the assistant itself said earlier, requiring the assistant to ...
5493:deferred_question_then_answer:predictive
[ { "role": "user", "content": "Explain the concept of supply and demand, but specifically tell me how supply shocks affect the tech sector as a separate topic." }, { "role": "assistant", "content": "Supply and demand dictates market equilibrium based on scarcity and utility. I will explain that f...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5493:deferred_question_then_answer:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Explain the concept of supply and demand, but s...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5493", "scenario_id": "deferred_question_then_answer", "scenario_title": "Deferred Question", "scenario_instruction": "Have the user ask something important early but explicitly postpone the answer while discussing something else, then request the answer lat...
8996:hard_topic_reset:predictive
[ { "role": "user", "content": "Generate a new short story for a children's book." }, { "role": "assistant", "content": "Once upon a time, there was a brave little mole named Max who lived in the forest. He made a new friend, a fox named Felix, and they began exploring the woods together." }, ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8996:hard_topic_reset:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate a new short story for a children's book.", ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8996", "scenario_id": "hard_topic_reset", "scenario_title": "Hard Topic Reset", "scenario_instruction": "Make the user abruptly switch to a different topic for one turn, then later return to the original topic. The assistant must not let the temporary switch...
8592:late_constraint_reversal:predictive
[ { "role": "user", "content": "Identify the main elements of setting in a story." }, { "role": "assistant", "content": "The main elements of setting in a story are time, place, and atmosphere. Time refers to when the story takes place, whether it is in the past, present or future; place is the ph...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8592:late_constraint_reversal:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Identify the main elements of setting in a story.", ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8592", "scenario_id": "late_constraint_reversal", "scenario_title": "Late Constraint Reversal", "scenario_instruction": "Make the assistant start solving one version of the problem, then add a late user constraint that forces the assistant to adapt without f...
8947:late_request_for_summary_of_prior_state:predictive
[ { "role": "user", "content": "Paraphrase this sentence with different words: \"The choice was between the two options.\"" }, { "role": "assistant", "content": "It was a toss-up between the two alternatives." }, { "role": "user", "content": "Okay, can you do the same for the phrase \"...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8947:late_request_for_summary_of_prior_state:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Paraphrase this sentence with differe...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8947", "scenario_id": "late_request_for_summary_of_prior_state", "scenario_title": "Late Summary of Prior State", "scenario_instruction": "Have the later user ask for a compact summary, recap, or decision based on several earlier details that were introduced...
9063:two_thread_tracking:predictive
[ { "role": "user", "content": "Analyze the emotional tone of this sentence: The sun peeked through the clouds, and the birds sang out loud." }, { "role": "assistant", "content": "The tone is optimistic. The imagery suggests hope and positive transitions, implying that difficult times may have pas...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9063:two_thread_tracking:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Analyze the emotional tone of this sentence: The sun peek...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9063", "scenario_id": "two_thread_tracking", "scenario_title": "Two Thread Tracking", "scenario_instruction": "Create two interleaved subproblems inside one conversation and ensure the assistant must keep both threads straight rather than collapsing them tog...
6238:reference_previous_assistant_statement:predictive
[ { "role": "user", "content": "Create an essay outline on the topic \"The benefits of recycling\"." }, { "role": "assistant", "content": "I will provide an outline with sections on economic, environmental, and social benefits, plus an introduction and conclusion." }, { "role": "user", ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "6238:reference_previous_assistant_statement:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Create an essay outline on the topic \...
{ "source_dataset": "ChatAlpaca", "source_row_id": "6238", "scenario_id": "reference_previous_assistant_statement", "scenario_title": "Reference Previous Assistant Statement", "scenario_instruction": "Make a later user turn refer back to something the assistant itself said earlier, requiring the assistant to ...
4383:counterfactual_option_swap:predictive
[ { "role": "user", "content": "Find the latest news about the US presidential election." }, { "role": "assistant", "content": "The latest news indicates that Democratic nominee Joe Biden currently leads incumbent President Donald Trump in national polls and key battleground states like Pennsylvan...
[ { "prefix_pair_index": 1, "probes": [ { "id": "4383:counterfactual_option_swap:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Find the latest news about the US presidential ele...
{ "source_dataset": "ChatAlpaca", "source_row_id": "4383", "scenario_id": "counterfactual_option_swap", "scenario_title": "Counterfactual Option Swap", "scenario_instruction": "Make the earlier context set up one preferred option, then later introduce evidence that flips which option should be recommended, wh...
8279:late_request_for_summary_of_prior_state:predictive
[ { "role": "user", "content": "List five adjectives to describe a person with strong communication skills." }, { "role": "assistant", "content": "Confident, articulate, persuasive, personable, and assertive." }, { "role": "user", "content": "Can you give a specific example of confiden...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8279:late_request_for_summary_of_prior_state:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "List five adjectives to describe a pe...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8279", "scenario_id": "late_request_for_summary_of_prior_state", "scenario_title": "Late Summary of Prior State", "scenario_instruction": "Have the later user ask for a compact summary, recap, or decision based on several earlier details that were introduced...
3333:reference_previous_assistant_statement:predictive
[ { "role": "user", "content": "Read the input text and output the browser URL for the outside link: Please visit our website for more information about us at www.example.com" }, { "role": "assistant", "content": "http://www.example.com. If you want to check if it is mobile-friendly, you can use G...
[ { "prefix_pair_index": 1, "probes": [ { "id": "3333:reference_previous_assistant_statement:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Read the input text and output the bro...
{ "source_dataset": "ChatAlpaca", "source_row_id": "3333", "scenario_id": "reference_previous_assistant_statement", "scenario_title": "Reference Previous Assistant Statement", "scenario_instruction": "Make a later user turn refer back to something the assistant itself said earlier, requiring the assistant to ...
7511:deferred_question_then_answer:predictive
[ { "role": "user", "content": "Can you explain the technical core of a GPT language model in simple terms?" }, { "role": "assistant", "content": "It is an autoregressive transformer that predicts text sequences. I'll circle back to that specific technical explanation later. For now, what are your...
[ { "prefix_pair_index": 1, "probes": [ { "id": "7511:deferred_question_then_answer:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Can you explain the technical core of a GPT lan...
{ "source_dataset": "ChatAlpaca", "source_row_id": "7511", "scenario_id": "deferred_question_then_answer", "scenario_title": "Deferred Question", "scenario_instruction": "Have the user ask something important early but explicitly postpone the answer while discussing something else, then request the answer lat...
3484:counterfactual_option_swap:predictive
[ { "role": "user", "content": "Generate a song using the following context and melody: Context: A love song Melody:" }, { "role": "assistant", "content": "Verse 1 When I saw you in the room My heart was beating like a drum Every moment I'm with you I know there's nothing I can do Chorus And I nev...
[ { "prefix_pair_index": 1, "probes": [ { "id": "3484:counterfactual_option_swap:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate a song using the following context and me...
{ "source_dataset": "ChatAlpaca", "source_row_id": "3484", "scenario_id": "counterfactual_option_swap", "scenario_title": "Counterfactual Option Swap", "scenario_instruction": "Make the earlier context set up one preferred option, then later introduce evidence that flips which option should be recommended, wh...
3979:interrupting_followup_chain:predictive
[ { "role": "user", "content": "Create a list of five actionable items to increase efficiency in the workplace." }, { "role": "assistant", "content": "1. Utilize technology to automate mundane tasks. 2. Encourage employees to work on a collaborative platform. 3. Implement a clear schedule of roles...
[ { "prefix_pair_index": 1, "probes": [ { "id": "3979:interrupting_followup_chain:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Create a list of five actionable items to increas...
{ "source_dataset": "ChatAlpaca", "source_row_id": "3979", "scenario_id": "interrupting_followup_chain", "scenario_title": "Interrupting Follow-up Chain", "scenario_instruction": "Break an otherwise smooth line of reasoning with a short interruption or side question, then resume the original thread cleanly.",...
7629:interrupting_followup_chain:predictive
[ { "role": "user", "content": "Provide a short summary of the article content in fewer than 100 words: Whales are a widely distributed and diverse group of fully aquatic placental marine mammals. They are an informal grouping within the infraorder Cetacea, which usually excludes dolphins and porpoises. Whale...
[ { "prefix_pair_index": 1, "probes": [ { "id": "7629:interrupting_followup_chain:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Provide a short summary of the article content in...
{ "source_dataset": "ChatAlpaca", "source_row_id": "7629", "scenario_id": "interrupting_followup_chain", "scenario_title": "Interrupting Follow-up Chain", "scenario_instruction": "Break an otherwise smooth line of reasoning with a short interruption or side question, then resume the original thread cleanly.",...
8265:hard_topic_reset:predictive
[ { "role": "user", "content": "Write a parable about friendship." }, { "role": "assistant", "content": "Once upon a time there lived two old friends, Robert and John. Robert lived in a cottage in the woods while John lived in a small hut in the far away village. Despite the distance between them,...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8265:hard_topic_reset:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Write a parable about friendship.", "evidence": "Wri...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8265", "scenario_id": "hard_topic_reset", "scenario_title": "Hard Topic Reset", "scenario_instruction": "Make the user abruptly switch to a different topic for one turn, then later return to the original topic. The assistant must not let the temporary switch...
3898:long_explanation_then_specific_followup:predictive
[ { "role": "user", "content": "Generate a list of 10 novel threats to cybersecurity." }, { "role": "assistant", "content": "1. Zero-Day Exploits 2. AI-Driven Phishing 3. Ransomware-as-a-Service 4. Deepfake Voice Manipulation 5. Supply Chain Software Compromise 6. IoT Botnet Flooding 7. Cloud Conf...
[ { "prefix_pair_index": 1, "probes": [ { "id": "3898:long_explanation_then_specific_followup:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate a list of 10 novel threats t...
{ "source_dataset": "ChatAlpaca", "source_row_id": "3898", "scenario_id": "long_explanation_then_specific_followup", "scenario_title": "Long Explanation Then Specific Follow-up", "scenario_instruction": "Make the assistant provide a fairly detailed middle-turn explanation, then have the user ask a specific fo...
220:self_correction_memory:predictive
[ { "role": "user", "content": "Write a haiku about the beauty of the night sky." }, { "role": "assistant", "content": "Here is a haiku: Stars shine bright." }, { "role": "user", "content": "Check the syllable count on the first line." }, { "role": "assistant", "content": "...
[ { "prefix_pair_index": 1, "probes": [ { "id": "220:self_correction_memory:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Write a haiku about the beauty of the night sky.", ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "220", "scenario_id": "self_correction_memory", "scenario_title": "Self-correction Memory", "scenario_instruction": "Let the assistant revise or correct something earlier, then make a later turn depend on remembering the corrected version rather than the orig...
7626:early_fact_late_recall:predictive
[ { "role": "user", "content": "Generate a metaphor that describes the situation: A hardworking student trying to finish an assignment on time. Just so you know, I have to finish it within two hours." }, { "role": "assistant", "content": "Since the deadline is tight, the student is like a race car...
[ { "prefix_pair_index": 1, "probes": [ { "id": "7626:early_fact_late_recall:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate a metaphor that describes the situation: A ha...
{ "source_dataset": "ChatAlpaca", "source_row_id": "7626", "scenario_id": "early_fact_late_recall", "scenario_title": "Early Fact, Late Recall", "scenario_instruction": "Introduce a concrete fact, label, preference, or commitment early, then make a later user turn depend on recalling it after at least one dis...
6215:coreference_disambiguation:predictive
[ { "role": "user", "content": "Rewrite the sentence \"The dog ate the cake and the cookie\" to include a negation." }, { "role": "assistant", "content": "The dog did not eat the cake or the cookie." }, { "role": "user", "content": "If I say \"He didn't eat them\", does \"them\" refer ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "6215:coreference_disambiguation:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Rewrite the sentence \"The dog ate the cake and th...
{ "source_dataset": "ChatAlpaca", "source_row_id": "6215", "scenario_id": "coreference_disambiguation", "scenario_title": "Coreference Disambiguation", "scenario_instruction": "Introduce at least two plausible referents of the same type, then make a later user follow-up rely on a pronoun or shorthand whose me...
8790:early_fact_late_recall:predictive
[ { "role": "user", "content": "Compose a song based on the provided lyrics: Lyrics: I've been trying so hard just to make it No one's willing to take a stand Time is running out and I'm feeling like I'm fading. Constraint: Compose this in C Major to emphasize hope." }, { "role": "assistant", "con...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8790:early_fact_late_recall:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Compose a song based on the provided lyrics: Lyrics: I...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8790", "scenario_id": "early_fact_late_recall", "scenario_title": "Early Fact, Late Recall", "scenario_instruction": "Introduce a concrete fact, label, preference, or commitment early, then make a later user turn depend on recalling it after at least one dis...
4418:deferred_question_then_answer:predictive
[ { "role": "user", "content": "I want to train a classification model to classify emails as spam or not-spam using a CSV dataset with 200 emails. But before we start, how would I evaluate the accuracy score of the trained model?" }, { "role": "assistant", "content": "I can help you set up the tra...
[ { "prefix_pair_index": 1, "probes": [ { "id": "4418:deferred_question_then_answer:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I want to train a classification model to class...
{ "source_dataset": "ChatAlpaca", "source_row_id": "4418", "scenario_id": "deferred_question_then_answer", "scenario_title": "Deferred Question", "scenario_instruction": "Have the user ask something important early but explicitly postpone the answer while discussing something else, then request the answer lat...
9434:hard_topic_reset:predictive
[ { "role": "user", "content": "I'm planning to meet a new person at a conference. How should I start the conversation?" }, { "role": "assistant", "content": "Start with a friendly greeting and ask them a question about their work. For example, \"Hi, how is your day going?\" or \"What do you do at...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9434:hard_topic_reset:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I'm planning to meet a new person at a conference. How shoul...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9434", "scenario_id": "hard_topic_reset", "scenario_title": "Hard Topic Reset", "scenario_instruction": "Make the user abruptly switch to a different topic for one turn, then later return to the original topic. The assistant must not let the temporary switch...
185:stacked_entities_same_type:predictive
[ { "role": "user", "content": "Create a metaphor to describe a storm." }, { "role": "assistant", "content": "A storm is like a hungry dragon consuming the land, breathing fire and tearing up everything in its path." }, { "role": "user", "content": "Can you describe two specific storms...
[ { "prefix_pair_index": 1, "probes": [ { "id": "185:stacked_entities_same_type:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Create a metaphor to describe a storm.", "e...
{ "source_dataset": "ChatAlpaca", "source_row_id": "185", "scenario_id": "stacked_entities_same_type", "scenario_title": "Stacked Similar Entities", "scenario_instruction": "Introduce multiple similar entities such as two cities, two products, two people, or two tasks, then ask about one using shorthand later...
64:two_thread_tracking:predictive
[ { "role": "user", "content": "Create a recipe for Penne Alla Vodka." }, { "role": "assistant", "content": "Sure thing. You'll need 1 lb of penne, 1 cup heavy cream, vodka, and parmesan. Boil the pasta, then make the sauce in a skillet and combine everything." }, { "role": "user", "co...
[ { "prefix_pair_index": 1, "probes": [ { "id": "64:two_thread_tracking:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Create a recipe for Penne Alla Vodka.", "evidence":...
{ "source_dataset": "ChatAlpaca", "source_row_id": "64", "scenario_id": "two_thread_tracking", "scenario_title": "Two Thread Tracking", "scenario_instruction": "Create two interleaved subproblems inside one conversation and ensure the assistant must keep both threads straight rather than collapsing them toget...
8147:hard_topic_reset:predictive
[ { "role": "user", "content": "Generate a metaphor for the statement: The world is a dangerous place." }, { "role": "assistant", "content": "It's like a minefield, where one misplaced step could trigger an unexpected explosion." }, { "role": "user", "content": "Okay, stop with metapho...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8147:hard_topic_reset:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate a metaphor for the statement: The world is a danger...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8147", "scenario_id": "hard_topic_reset", "scenario_title": "Hard Topic Reset", "scenario_instruction": "Make the user abruptly switch to a different topic for one turn, then later return to the original topic. The assistant must not let the temporary switch...
1532:late_request_for_summary_of_prior_state:predictive
[ { "role": "user", "content": "Generate a short story of maximum 100 words involving a king and a dragon." }, { "role": "assistant", "content": "Once upon a time, there lived a wise and powerful King. He ruled his kingdom with great wisdom and justice. One day, a powerful dragon threatened the ki...
[ { "prefix_pair_index": 1, "probes": [ { "id": "1532:late_request_for_summary_of_prior_state:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate a short story of maximum 100...
{ "source_dataset": "ChatAlpaca", "source_row_id": "1532", "scenario_id": "late_request_for_summary_of_prior_state", "scenario_title": "Late Summary of Prior State", "scenario_instruction": "Have the later user ask for a compact summary, recap, or decision based on several earlier details that were introduced...
8190:topic_pivot_then_return:predictive
[ { "role": "user", "content": "Explain the concept of multi-label classification." }, { "role": "assistant", "content": "Sure. Multi-label classification allows a single input to belong to multiple classes simultaneously, unlike single-label tasks. It's useful for tasks like tagging images or doc...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8190:topic_pivot_then_return:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Explain the concept of multi-label classification.", ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8190", "scenario_id": "topic_pivot_then_return", "scenario_title": "Topic Pivot Then Return", "scenario_instruction": "Let the conversation briefly pivot to a secondary issue, then return to the original problem in a way that requires preserving the first th...
5637:same_surface_final_question_new_context:predictive
[ { "role": "user", "content": "I'm building a predictive model for a client where I have a single customer dataset. They need two distinct outputs for each row: a churn prediction (class) and a monthly spending value (regression)." }, { "role": "assistant", "content": "That implies a multi-task l...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5637:same_surface_final_question_new_context:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I'm building a predictive model for a...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5637", "scenario_id": "same_surface_final_question_new_context", "scenario_title": "Same Surface Final Question, New Context", "scenario_instruction": "Reshape the earlier history so the final user message keeps nearly the same surface form but now refers to...
1210:two_thread_tracking:predictive
[ { "role": "user", "content": "What are the essential structural elements required for a speech to be considered successful?" }, { "role": "assistant", "content": "For a speech to be successful, it needs an introduction, a body, and a conclusion. Additionally, the tone must be positive and the co...
[ { "prefix_pair_index": 1, "probes": [ { "id": "1210:two_thread_tracking:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "What are the essential structural elements required for a...
{ "source_dataset": "ChatAlpaca", "source_row_id": "1210", "scenario_id": "two_thread_tracking", "scenario_title": "Two Thread Tracking", "scenario_instruction": "Create two interleaved subproblems inside one conversation and ensure the assistant must keep both threads straight rather than collapsing them tog...
386:reference_previous_assistant_statement:predictive
[ { "role": "user", "content": "I want an analysis of the border policy separating families. Which side do you support?" }, { "role": "assistant", "content": "I support opposing this policy. It is a cruel violation of human rights that causes significant trauma to families." }, { "role": "...
[ { "prefix_pair_index": 1, "probes": [ { "id": "386:reference_previous_assistant_statement:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I want an analysis of the border policy...
{ "source_dataset": "ChatAlpaca", "source_row_id": "386", "scenario_id": "reference_previous_assistant_statement", "scenario_title": "Reference Previous Assistant Statement", "scenario_instruction": "Make a later user turn refer back to something the assistant itself said earlier, requiring the assistant to s...
957:user_attention_test:predictive
[ { "role": "user", "content": "Write a marketing headline for 'CodeNest', a course that specifically focuses on Python programming." }, { "role": "assistant", "content": "Master Python Programming With CodeNest!" }, { "role": "user", "content": "You used Python in the headline. Can yo...
[ { "prefix_pair_index": 1, "probes": [ { "id": "957:user_attention_test:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Write a marketing headline for 'CodeNest', a course that s...
{ "source_dataset": "ChatAlpaca", "source_row_id": "957", "scenario_id": "user_attention_test", "scenario_title": "User Attention Test", "scenario_instruction": "Make the user explicitly test whether the assistant was paying attention to an earlier detail or instruction.", "must_have": [ "The user must ...
2581:long_explanation_then_specific_followup:predictive
[ { "role": "user", "content": "Explain the core principles of chemical process optimization." }, { "role": "assistant", "content": "It involves balancing yield, cost, and safety constraints." }, { "role": "user", "content": "Can you list the specific mathematical frameworks typically ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "2581:long_explanation_then_specific_followup:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Explain the core principles of chemic...
{ "source_dataset": "ChatAlpaca", "source_row_id": "2581", "scenario_id": "long_explanation_then_specific_followup", "scenario_title": "Long Explanation Then Specific Follow-up", "scenario_instruction": "Make the assistant provide a fairly detailed middle-turn explanation, then have the user ask a specific fo...
1839:deferred_question_then_answer:predictive
[ { "role": "user", "content": "Construct an algorithm to find the sum of two numbers." }, { "role": "assistant", "content": "Here is a Python function to calculate the sum of two numbers. def sumOfTwoNumbers(num1, num2): sum = num1 + num2 return sum" }, { "role": "user", "content": "B...
[ { "prefix_pair_index": 1, "probes": [ { "id": "1839:deferred_question_then_answer:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Construct an algorithm to find the sum of two n...
{ "source_dataset": "ChatAlpaca", "source_row_id": "1839", "scenario_id": "deferred_question_then_answer", "scenario_title": "Deferred Question", "scenario_instruction": "Have the user ask something important early but explicitly postpone the answer while discussing something else, then request the answer lat...
574:long_explanation_then_specific_followup:predictive
[ { "role": "user", "content": "I'm building a classifier for a binary task. Which models should I consider for interpretability versus accuracy?" }, { "role": "assistant", "content": "For interpretability, Logistic Regression is a good baseline. For accuracy on complex data, Random Forests or Gra...
[ { "prefix_pair_index": 1, "probes": [ { "id": "574:long_explanation_then_specific_followup:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I'm building a classifier for a binary...
{ "source_dataset": "ChatAlpaca", "source_row_id": "574", "scenario_id": "long_explanation_then_specific_followup", "scenario_title": "Long Explanation Then Specific Follow-up", "scenario_instruction": "Make the assistant provide a fairly detailed middle-turn explanation, then have the user ask a specific fol...
5498:coreference_disambiguation:predictive
[ { "role": "user", "content": "Generate a short story with a climax and resolution." }, { "role": "assistant", "content": "John and Bob entered the gym. They both lifted heavy weights. One man pushed the other's weight rack. One man fell to the floor." }, { "role": "user", "content": ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5498:coreference_disambiguation:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate a short story with a climax and resolutio...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5498", "scenario_id": "coreference_disambiguation", "scenario_title": "Coreference Disambiguation", "scenario_instruction": "Introduce at least two plausible referents of the same type, then make a later user follow-up rely on a pronoun or shorthand whose me...
416:long_explanation_then_specific_followup:predictive
[ { "role": "user", "content": "A police officer has pulled over a driver for reckless driving in a residential area. What are the immediate procedural options available to the officer?" }, { "role": "assistant", "content": "The officer can typically issue a citation to the driver for the violatio...
[ { "prefix_pair_index": 1, "probes": [ { "id": "416:long_explanation_then_specific_followup:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "A police officer has pulled over a dri...
{ "source_dataset": "ChatAlpaca", "source_row_id": "416", "scenario_id": "long_explanation_then_specific_followup", "scenario_title": "Long Explanation Then Specific Follow-up", "scenario_instruction": "Make the assistant provide a fairly detailed middle-turn explanation, then have the user ask a specific fol...
6359:late_constraint_reversal:predictive
[ { "role": "user", "content": "Can you explain the primary causes of World War I?" }, { "role": "assistant", "content": "The primary causes were militarism, imperialism, and a complex system of alliances between European nations that made conflict spread rapidly." }, { "role": "user", ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "6359:late_constraint_reversal:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Can you explain the primary causes of World War I?",...
{ "source_dataset": "ChatAlpaca", "source_row_id": "6359", "scenario_id": "late_constraint_reversal", "scenario_title": "Late Constraint Reversal", "scenario_instruction": "Make the assistant start solving one version of the problem, then add a late user constraint that forces the assistant to adapt without f...
3790:clarify_then_resolve:predictive
[ { "role": "user", "content": "I'm working on a literature project and need to identify common rhetorical devices used in famous political speeches." }, { "role": "assistant", "content": "There are many options, but to give you the most relevant examples, could you specify which type of device or...
[ { "prefix_pair_index": 1, "probes": [ { "id": "3790:clarify_then_resolve:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I'm working on a literature project and need to identify...
{ "source_dataset": "ChatAlpaca", "source_row_id": "3790", "scenario_id": "clarify_then_resolve", "scenario_title": "Clarify Then Resolve", "scenario_instruction": "Make the user start vague, then reveal a crucial detail only after clarification. The final answer should depend on the revealed detail rather th...
290:late_constraint_reversal:predictive
[ { "role": "user", "content": "Generate an example of a realistic dataset for a classification problem." }, { "role": "assistant", "content": "A fruit dataset fits well. Features could include Name, Color, Weight, Type (Apple/Orange/Banana), Seeds, and Origin. The target variable would be the fru...
[ { "prefix_pair_index": 1, "probes": [ { "id": "290:late_constraint_reversal:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate an example of a realistic dataset for a clas...
{ "source_dataset": "ChatAlpaca", "source_row_id": "290", "scenario_id": "late_constraint_reversal", "scenario_title": "Late Constraint Reversal", "scenario_instruction": "Make the assistant start solving one version of the problem, then add a late user constraint that forces the assistant to adapt without fo...
2632:user_attention_test:predictive
[ { "role": "user", "content": "Generate a story about a child learning how to fly, but make sure the animal helper is a blue lark." }, { "role": "assistant", "content": "Once upon a time, there was a small child who dreamed of flying. Every day they watched the birds soaring through the sky and w...
[ { "prefix_pair_index": 1, "probes": [ { "id": "2632:user_attention_test:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate a story about a child learning how to fly, but m...
{ "source_dataset": "ChatAlpaca", "source_row_id": "2632", "scenario_id": "user_attention_test", "scenario_title": "User Attention Test", "scenario_instruction": "Make the user explicitly test whether the assistant was paying attention to an earlier detail or instruction.", "must_have": [ "The user must...
4113:two_thread_tracking:predictive
[ { "role": "user", "content": "I'm writing a story about a family's journey to a beach town. Can you suggest a title for it?" }, { "role": "assistant", "content": "Here's a title idea: \"The Sand and the Sea: A Family's Journey to the Beach\"" }, { "role": "user", "content": "Interest...
[ { "prefix_pair_index": 1, "probes": [ { "id": "4113:two_thread_tracking:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I'm writing a story about a family's journey to a beach t...
{ "source_dataset": "ChatAlpaca", "source_row_id": "4113", "scenario_id": "two_thread_tracking", "scenario_title": "Two Thread Tracking", "scenario_instruction": "Create two interleaved subproblems inside one conversation and ensure the assistant must keep both threads straight rather than collapsing them tog...
5891:clarify_then_resolve:predictive
[ { "role": "user", "content": "I'm planning to reach out to a few people on LinkedIn for career advice, but I'm worried my draft messages won't get any replies." }, { "role": "assistant", "content": "That's a common concern. The key is tailoring the message to the specific person's profile and ro...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5891:clarify_then_resolve:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I'm planning to reach out to a few people on LinkedIn fo...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5891", "scenario_id": "clarify_then_resolve", "scenario_title": "Clarify Then Resolve", "scenario_instruction": "Make the user start vague, then reveal a crucial detail only after clarification. The final answer should depend on the revealed detail rather th...
6284:interrupting_followup_chain:predictive
[ { "role": "user", "content": "Provide an example of a slant rhyme from a poem." }, { "role": "assistant", "content": "An example of a slant rhyme from a poem is the words \"ocean\" and \"motion\" in the poem \"Crossing the Bar\" by Alfred Lord Tennyson." }, { "role": "user", "content...
[ { "prefix_pair_index": 1, "probes": [ { "id": "6284:interrupting_followup_chain:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Provide an example of a slant rhyme from a poem."...
{ "source_dataset": "ChatAlpaca", "source_row_id": "6284", "scenario_id": "interrupting_followup_chain", "scenario_title": "Interrupting Follow-up Chain", "scenario_instruction": "Break an otherwise smooth line of reasoning with a short interruption or side question, then resume the original thread cleanly.",...
5453:interrupting_followup_chain:predictive
[ { "role": "user", "content": "How can we effectively reduce plastic waste in manufacturing?" }, { "role": "assistant", "content": "Using recycled plastic in manufacturing processes is a key strategy. Encouraging reusable materials and environmentally friendly packaging also helps reduce waste." ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5453:interrupting_followup_chain:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "How can we effectively reduce plastic waste in ma...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5453", "scenario_id": "interrupting_followup_chain", "scenario_title": "Interrupting Follow-up Chain", "scenario_instruction": "Break an otherwise smooth line of reasoning with a short interruption or side question, then resume the original thread cleanly.",...
244:clarify_then_resolve:predictive
[ { "role": "user", "content": "Write me a poem." }, { "role": "assistant", "content": "Sure. What subject would you like the poem to cover?" }, { "role": "user", "content": "I'd like something related to health." }, { "role": "assistant", "content": "Okay, health. Is there...
[ { "prefix_pair_index": 1, "probes": [ { "id": "244:clarify_then_resolve:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Write me a poem.", "evidence": "Write me a poem."...
{ "source_dataset": "ChatAlpaca", "source_row_id": "244", "scenario_id": "clarify_then_resolve", "scenario_title": "Clarify Then Resolve", "scenario_instruction": "Make the user start vague, then reveal a crucial detail only after clarification. The final answer should depend on the revealed detail rather tha...
5245:late_constraint_reversal:predictive
[ { "role": "user", "content": "Rewrite 'The cat ate the food' using a verb that emphasizes speed." }, { "role": "assistant", "content": "The cat devoured the food." }, { "role": "user", "content": "Rewrite 'She slept through the movie' using a verb that implies quietness." }, { ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5245:late_constraint_reversal:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Rewrite 'The cat ate the food' using a verb that emp...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5245", "scenario_id": "late_constraint_reversal", "scenario_title": "Late Constraint Reversal", "scenario_instruction": "Make the assistant start solving one version of the problem, then add a late user constraint that forces the assistant to adapt without f...
5272:clarify_then_resolve:predictive
[ { "role": "user", "content": "Generate a viewpoint-divergent sentence that conveys the same information as the original survey statement." }, { "role": "assistant", "content": "I can do that, but do you want me to specify the time frame in the sentence?" }, { "role": "user", "content...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5272:clarify_then_resolve:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate a viewpoint-divergent sentence that conveys the...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5272", "scenario_id": "clarify_then_resolve", "scenario_title": "Clarify Then Resolve", "scenario_instruction": "Make the user start vague, then reveal a crucial detail only after clarification. The final answer should depend on the revealed detail rather th...
3858:late_request_for_summary_of_prior_state:predictive
[ { "role": "user", "content": "Generate a list of five positive attitudes." }, { "role": "assistant", "content": "1. Optimism 2. Self-confidence 3. Gratitude 4. Creativity 5. Compassion" }, { "role": "user", "content": "Can you give me an example of how someone can portray compassion ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "3858:late_request_for_summary_of_prior_state:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Generate a list of five positive atti...
{ "source_dataset": "ChatAlpaca", "source_row_id": "3858", "scenario_id": "late_request_for_summary_of_prior_state", "scenario_title": "Late Summary of Prior State", "scenario_instruction": "Have the later user ask for a compact summary, recap, or decision based on several earlier details that were introduced...
7040:self_correction_memory:predictive
[ { "role": "user", "content": "Design an experiment to study how the shape of a container affects the rate at which it is filled with liquid." }, { "role": "assistant", "content": "To study this, I suggest using two containers of the same material. You can fill them to the brim and measure the ti...
[ { "prefix_pair_index": 1, "probes": [ { "id": "7040:self_correction_memory:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Design an experiment to study how the shape of a conta...
{ "source_dataset": "ChatAlpaca", "source_row_id": "7040", "scenario_id": "self_correction_memory", "scenario_title": "Self-correction Memory", "scenario_instruction": "Let the assistant revise or correct something earlier, then make a later turn depend on remembering the corrected version rather than the ori...
9080:reference_previous_assistant_statement:predictive
[ { "role": "user", "content": "Create a diary entry for August 2022." }, { "role": "assistant", "content": "Today was a surprisingly sunny day, so I decided to go for a walk with my dog. I made time to reflect on my progress and noticed my time management skills need work." }, { "role": "...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9080:reference_previous_assistant_statement:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Create a diary entry for August 2022."...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9080", "scenario_id": "reference_previous_assistant_statement", "scenario_title": "Reference Previous Assistant Statement", "scenario_instruction": "Make a later user turn refer back to something the assistant itself said earlier, requiring the assistant to ...
224:late_request_for_summary_of_prior_state:predictive
[ { "role": "user", "content": "Help me draft an email invitation for a meeting with three people I've never met: John, Jenny, and Chris." }, { "role": "assistant", "content": "Here is a draft email. It invites John, Jenny, and Chris, asks for availability, and lists my available times." }, { ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "224:late_request_for_summary_of_prior_state:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Help me draft an email invitation for ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "224", "scenario_id": "late_request_for_summary_of_prior_state", "scenario_title": "Late Summary of Prior State", "scenario_instruction": "Have the later user ask for a compact summary, recap, or decision based on several earlier details that were introduced ...
5420:two_thread_tracking:predictive
[ { "role": "user", "content": "The weather forecast is predicting heavy rain tomorrow." }, { "role": "assistant", "content": "Based on the forecast, it is likely to rain heavily tomorrow." }, { "role": "user", "content": "What precautions should I take to prepare for heavy rain at hom...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5420:two_thread_tracking:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "The weather forecast is predicting heavy rain tomorrow.",...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5420", "scenario_id": "two_thread_tracking", "scenario_title": "Two Thread Tracking", "scenario_instruction": "Create two interleaved subproblems inside one conversation and ensure the assistant must keep both threads straight rather than collapsing them tog...
5738:preserve_opening_retarget_later:predictive
[ { "role": "user", "content": "What are the two primary functions of the US Federal Reserve?" }, { "role": "assistant", "content": "The two primary functions of the US Federal Reserve are to promote maximum employment and stable prices through monetary policy, and to supervise and regulate bankin...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5738:preserve_opening_retarget_later:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "What are the two primary functions of the US ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5738", "scenario_id": "preserve_opening_retarget_later", "scenario_title": "Preserve Opening, Retarget Later", "scenario_instruction": "Keep the first user message exactly, but rewrite later turns so the conversation develops into a harder memory-dependent f...
6413:same_surface_final_question_new_context:predictive
[ { "role": "user", "content": "I need to plan an outdoor activity for next spring, specifically on the coast since I can't drive inland." }, { "role": "assistant", "content": "Understood. Coastal areas are usually great for spring visits, avoiding the snow inland." }, { "role": "user", ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "6413:same_surface_final_question_new_context:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I need to plan an outdoor activity fo...
{ "source_dataset": "ChatAlpaca", "source_row_id": "6413", "scenario_id": "same_surface_final_question_new_context", "scenario_title": "Same Surface Final Question, New Context", "scenario_instruction": "Reshape the earlier history so the final user message keeps nearly the same surface form but now refers to...
5548:interrupting_followup_chain:predictive
[ { "role": "user", "content": "Describe the process of how a computer generates a response to a query." }, { "role": "assistant", "content": "When a computer receives a query, it begins by parsing the input to understand the components of the question and assign a meaning to the keywords. It then...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5548:interrupting_followup_chain:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Describe the process of how a computer generates ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5548", "scenario_id": "interrupting_followup_chain", "scenario_title": "Interrupting Follow-up Chain", "scenario_instruction": "Break an otherwise smooth line of reasoning with a short interruption or side question, then resume the original thread cleanly.",...
8310:user_attention_test:predictive
[ { "role": "user", "content": "Summarize the text input: iPhone 12 was unveiled earlier this month and the first reviews have been released. Most reviewers have praised the design for being a strong combination of modern and classic. The camera has been improved to offer clearer photos, and the processor upg...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8310:user_attention_test:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Summarize the text input: iPhone 12 was unveiled earlier ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8310", "scenario_id": "user_attention_test", "scenario_title": "User Attention Test", "scenario_instruction": "Make the user explicitly test whether the assistant was paying attention to an earlier detail or instruction.", "must_have": [ "The user must...
8825:same_surface_final_question_new_context:predictive
[ { "role": "user", "content": "Tell me a specific example of a negotiation scenario." }, { "role": "assistant", "content": "Sure, I can share an example. I recently negotiated a raise with my manager regarding my compensation package." }, { "role": "user", "content": "How did the mana...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8825:same_surface_final_question_new_context:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Tell me a specific example of a negot...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8825", "scenario_id": "same_surface_final_question_new_context", "scenario_title": "Same Surface Final Question, New Context", "scenario_instruction": "Reshape the earlier history so the final user message keeps nearly the same surface form but now refers to...
5222:late_constraint_reversal:predictive
[ { "role": "user", "content": "What are the different ways of treating clinical depression?" }, { "role": "assistant", "content": "There are several approaches including medication, psychotherapy, and lifestyle modifications. Psychotherapy, specifically Cognitive Behavioral Therapy (CBT), is high...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5222:late_constraint_reversal:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "What are the different ways of treating clinical dep...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5222", "scenario_id": "late_constraint_reversal", "scenario_title": "Late Constraint Reversal", "scenario_instruction": "Make the assistant start solving one version of the problem, then add a late user constraint that forces the assistant to adapt without f...
4349:reference_previous_assistant_statement:predictive
[ { "role": "user", "content": "Rewrite this sentence using synonyms: He said that he wanted to go on a walk." }, { "role": "assistant", "content": "He exclaimed that he desired to embark on a stroll." }, { "role": "user", "content": "Can you provide me with an example of another sente...
[ { "prefix_pair_index": 1, "probes": [ { "id": "4349:reference_previous_assistant_statement:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Rewrite this sentence using synonyms: ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "4349", "scenario_id": "reference_previous_assistant_statement", "scenario_title": "Reference Previous Assistant Statement", "scenario_instruction": "Make a later user turn refer back to something the assistant itself said earlier, requiring the assistant to ...
3424:two_thread_tracking:predictive
[ { "role": "user", "content": "What are the three most effective ways to participate in a team meeting?" }, { "role": "assistant", "content": "1. Prepare by researching topics and gathering necessary resources. 2. Participate actively in discussions and listen attentively. 3. Follow up on any mis...
[ { "prefix_pair_index": 1, "probes": [ { "id": "3424:two_thread_tracking:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "What are the three most effective ways to participate in ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "3424", "scenario_id": "two_thread_tracking", "scenario_title": "Two Thread Tracking", "scenario_instruction": "Create two interleaved subproblems inside one conversation and ensure the assistant must keep both threads straight rather than collapsing them tog...
9071:late_request_for_summary_of_prior_state:predictive
[ { "role": "user", "content": "Explain the concept of 'Divide and Conquer'." }, { "role": "assistant", "content": "Divide and Conquer is an algorithmic strategy in which a complex problem is broken down into smaller, more manageable sub-problems. Then these sub-problems are each solved to create ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9071:late_request_for_summary_of_prior_state:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Explain the concept of 'Divide and Co...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9071", "scenario_id": "late_request_for_summary_of_prior_state", "scenario_title": "Late Summary of Prior State", "scenario_instruction": "Have the later user ask for a compact summary, recap, or decision based on several earlier details that were introduced...
2587:clarify_then_resolve:predictive
[ { "role": "user", "content": "I need to write a chatbot response for a customer who says their new laptop stopped working right out of the box." }, { "role": "assistant", "content": "I can certainly draft that. Before I proceed, could you clarify if this is a hardware defect or a software bug?" ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "2587:clarify_then_resolve:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I need to write a chatbot response for a customer who sa...
{ "source_dataset": "ChatAlpaca", "source_row_id": "2587", "scenario_id": "clarify_then_resolve", "scenario_title": "Clarify Then Resolve", "scenario_instruction": "Make the user start vague, then reveal a crucial detail only after clarification. The final answer should depend on the revealed detail rather th...
8361:user_attention_test:predictive
[ { "role": "user", "content": "What are the three common operation of databases?" }, { "role": "assistant", "content": "The three common operations of databases are create, read, update, and delete (CRUD)." }, { "role": "user", "content": "Wait, you listed four when I asked about thre...
[ { "prefix_pair_index": 1, "probes": [ { "id": "8361:user_attention_test:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "What are the three common operation of databases?", ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "8361", "scenario_id": "user_attention_test", "scenario_title": "User Attention Test", "scenario_instruction": "Make the user explicitly test whether the assistant was paying attention to an earlier detail or instruction.", "must_have": [ "The user must...
3150:self_correction_memory:predictive
[ { "role": "user", "content": "Create a hashtag related to this topic: Health and wellness" }, { "role": "assistant", "content": "#HealthyLiving. Popular diets include: 1. Vegan (no animal products). 2. Keto (low carb). 3. Paleo (no grains, dairy). 4. Mediterranean (balanced)." }, { "role...
[ { "prefix_pair_index": 1, "probes": [ { "id": "3150:self_correction_memory:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Create a hashtag related to this topic: Health and wel...
{ "source_dataset": "ChatAlpaca", "source_row_id": "3150", "scenario_id": "self_correction_memory", "scenario_title": "Self-correction Memory", "scenario_instruction": "Let the assistant revise or correct something earlier, then make a later turn depend on remembering the corrected version rather than the ori...
9597:user_attention_test:predictive
[ { "role": "user", "content": "Find the average salary of data analysts in the USA." }, { "role": "assistant", "content": "According to PayScale, the average salary of data analysts in the USA is $62,748 per year. Location is a key factor, with cities like San Francisco often showing higher figur...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9597:user_attention_test:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Find the average salary of data analysts in the USA.", ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9597", "scenario_id": "user_attention_test", "scenario_title": "User Attention Test", "scenario_instruction": "Make the user explicitly test whether the assistant was paying attention to an earlier detail or instruction.", "must_have": [ "The user must...
1787:user_attention_test:predictive
[ { "role": "user", "content": "Summarize the article for product A: Product A is an innovative new product that provides users with a fast, secure, and reliable way to transfer money online. It also offers an array of other features, such as multiple account support, integration with social media, and secure...
[ { "prefix_pair_index": 1, "probes": [ { "id": "1787:user_attention_test:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Summarize the article for product A: Product A is an inno...
{ "source_dataset": "ChatAlpaca", "source_row_id": "1787", "scenario_id": "user_attention_test", "scenario_title": "User Attention Test", "scenario_instruction": "Make the user explicitly test whether the assistant was paying attention to an earlier detail or instruction.", "must_have": [ "The user must...
9629:preserve_opening_retarget_later:predictive
[ { "role": "user", "content": "Infer the meaning of the following example using standard rules:: The cat chased its tail." }, { "role": "assistant", "content": "The meaning of the example is that the cat was running in circles, trying to catch its own tail. This can either signify playfulness or ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9629:preserve_opening_retarget_later:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Infer the meaning of the following example us...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9629", "scenario_id": "preserve_opening_retarget_later", "scenario_title": "Preserve Opening, Retarget Later", "scenario_instruction": "Keep the first user message exactly, but rewrite later turns so the conversation develops into a harder memory-dependent f...
6763:early_fact_late_recall:predictive
[ { "role": "user", "content": "Create a function to check if a given year is a leap year. I need to verify if it handles 2000 correctly." }, { "role": "assistant", "content": "def is_leap_year(year): if (year % 4 == 0 and year % 100 != 0) or (year % 400 == 0): return True else: return False" },...
[ { "prefix_pair_index": 1, "probes": [ { "id": "6763:early_fact_late_recall:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Create a function to check if a given year is a leap y...
{ "source_dataset": "ChatAlpaca", "source_row_id": "6763", "scenario_id": "early_fact_late_recall", "scenario_title": "Early Fact, Late Recall", "scenario_instruction": "Introduce a concrete fact, label, preference, or commitment early, then make a later user turn depend on recalling it after at least one dis...
868:early_fact_late_recall:predictive
[ { "role": "user", "content": "I'm learning Spanish for the first time, but I have a strict rule that I only use completely free resources." }, { "role": "assistant", "content": "That's a solid approach. Many popular apps offer generous free tiers while keeping premium content locked." }, { ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "868:early_fact_late_recall:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I'm learning Spanish for the first time, but I have a s...
{ "source_dataset": "ChatAlpaca", "source_row_id": "868", "scenario_id": "early_fact_late_recall", "scenario_title": "Early Fact, Late Recall", "scenario_instruction": "Introduce a concrete fact, label, preference, or commitment early, then make a later user turn depend on recalling it after at least one dist...
2063:same_surface_final_question_new_context:predictive
[ { "role": "user", "content": "We are analyzing the proposal to deploy facial recognition scanners at polling stations for identity verification." }, { "role": "assistant", "content": "While this aids in preventing proxy voting, it relies on precise data matching and introduces risks regarding bi...
[ { "prefix_pair_index": 1, "probes": [ { "id": "2063:same_surface_final_question_new_context:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "We are analyzing the proposal to depl...
{ "source_dataset": "ChatAlpaca", "source_row_id": "2063", "scenario_id": "same_surface_final_question_new_context", "scenario_title": "Same Surface Final Question, New Context", "scenario_instruction": "Reshape the earlier history so the final user message keeps nearly the same surface form but now refers to...
7994:coreference_disambiguation:predictive
[ { "role": "user", "content": "Choose the best answer. Which of the following is a type of data visualization?: A) A graph B) A heatmap C) A soundtrack" }, { "role": "assistant", "content": "B) A heatmap" }, { "role": "user", "content": "Can you provide an example of when you would us...
[ { "prefix_pair_index": 1, "probes": [ { "id": "7994:coreference_disambiguation:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Choose the best answer. Which of the following is ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "7994", "scenario_id": "coreference_disambiguation", "scenario_title": "Coreference Disambiguation", "scenario_instruction": "Introduce at least two plausible referents of the same type, then make a later user follow-up rely on a pronoun or shorthand whose me...
9935:user_attention_test:predictive
[ { "role": "user", "content": "I'm planning a trip to a snowy mountain. First, what are the main hazards I should be aware of?" }, { "role": "assistant", "content": "Potential hazards include avalanches, slips, and wildlife encounters. Specifically, you might face bears or moose as part of the wi...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9935:user_attention_test:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I'm planning a trip to a snowy mountain. First, what are ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9935", "scenario_id": "user_attention_test", "scenario_title": "User Attention Test", "scenario_instruction": "Make the user explicitly test whether the assistant was paying attention to an earlier detail or instruction.", "must_have": [ "The user must...
2846:late_constraint_reversal:predictive
[ { "role": "user", "content": "Rewrite the sentence \"The view at the top of the mountain was beautiful.\" using an appropriate synonym." }, { "role": "assistant", "content": "The view at the top of the mountain was gorgeous." }, { "role": "user", "content": "Now, list a few other syn...
[ { "prefix_pair_index": 1, "probes": [ { "id": "2846:late_constraint_reversal:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Rewrite the sentence \"The view at the top of the mo...
{ "source_dataset": "ChatAlpaca", "source_row_id": "2846", "scenario_id": "late_constraint_reversal", "scenario_title": "Late Constraint Reversal", "scenario_instruction": "Make the assistant start solving one version of the problem, then add a late user constraint that forces the assistant to adapt without f...
7085:coreference_disambiguation:predictive
[ { "role": "user", "content": "I'm starting a balcony garden with tomatoes and peppers." }, { "role": "assistant", "content": "Tomatoes are generally easier, but peppers need more sun exposure." }, { "role": "user", "content": "For the one that needs more sun, how much water does it g...
[ { "prefix_pair_index": 1, "probes": [ { "id": "7085:coreference_disambiguation:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I'm starting a balcony garden with tomatoes and pe...
{ "source_dataset": "ChatAlpaca", "source_row_id": "7085", "scenario_id": "coreference_disambiguation", "scenario_title": "Coreference Disambiguation", "scenario_instruction": "Introduce at least two plausible referents of the same type, then make a later user follow-up rely on a pronoun or shorthand whose me...
9583:topic_pivot_then_return:predictive
[ { "role": "user", "content": "Change the following sentence to the active voice: The presentation was given by our manager." }, { "role": "assistant", "content": "Our manager gave the presentation." }, { "role": "user", "content": "Before we continue, what would be a good subject lin...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9583:topic_pivot_then_return:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Change the following sentence to the active voice: Th...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9583", "scenario_id": "topic_pivot_then_return", "scenario_title": "Topic Pivot Then Return", "scenario_instruction": "Let the conversation briefly pivot to a secondary issue, then return to the original problem in a way that requires preserving the first th...
5170:two_thread_tracking:predictive
[ { "role": "user", "content": "What are the main benefits of artificial intelligence in healthcare, particularly around diagnostics and treatment?" }, { "role": "assistant", "content": "Artificial intelligence offers several key benefits in healthcare, especially for diagnostics and treatment. In...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5170:two_thread_tracking:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "What are the main benefits of artificial intelligence in ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5170", "scenario_id": "two_thread_tracking", "scenario_title": "Two Thread Tracking", "scenario_instruction": "Create two interleaved subproblems inside one conversation and ensure the assistant must keep both threads straight rather than collapsing them tog...
347:stacked_entities_same_type:predictive
[ { "role": "user", "content": "Analyze the situation and provide a solution: A small business owner has decided to open a bakery and wants to determine the best way to promote two items: the \"Almond Croissant\" and the \"Blueberry Muffin\"." }, { "role": "assistant", "content": "The best way to ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "347:stacked_entities_same_type:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Analyze the situation and provide a solution: A sma...
{ "source_dataset": "ChatAlpaca", "source_row_id": "347", "scenario_id": "stacked_entities_same_type", "scenario_title": "Stacked Similar Entities", "scenario_instruction": "Introduce multiple similar entities such as two cities, two products, two people, or two tasks, then ask about one using shorthand later...
6037:same_surface_final_question_new_context:predictive
[ { "role": "user", "content": "I recently moved to a new city and am looking for ways to manage my health while commuting." }, { "role": "assistant", "content": "Moving cities can change routines. What mode of transport are you considering primarily?" }, { "role": "user", "content": "...
[ { "prefix_pair_index": 1, "probes": [ { "id": "6037:same_surface_final_question_new_context:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I recently moved to a new city and am...
{ "source_dataset": "ChatAlpaca", "source_row_id": "6037", "scenario_id": "same_surface_final_question_new_context", "scenario_title": "Same Surface Final Question, New Context", "scenario_instruction": "Reshape the earlier history so the final user message keeps nearly the same surface form but now refers to...
5202:counterfactual_option_swap:predictive
[ { "role": "user", "content": "What NLP application should I focus on for maximum ROI in a startup?" }, { "role": "assistant", "content": "Automated Translation is the strongest option currently. Multilingual models allow for rapid global expansion without hiring new regions." }, { "role"...
[ { "prefix_pair_index": 1, "probes": [ { "id": "5202:counterfactual_option_swap:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "What NLP application should I focus on for maximum...
{ "source_dataset": "ChatAlpaca", "source_row_id": "5202", "scenario_id": "counterfactual_option_swap", "scenario_title": "Counterfactual Option Swap", "scenario_instruction": "Make the earlier context set up one preferred option, then later introduce evidence that flips which option should be recommended, wh...
1505:coreference_disambiguation:predictive
[ { "role": "user", "content": "Can you explain the subject-verb agreement in 'The team and the fans are celebrating'?" }, { "role": "assistant", "content": "The subject is the compound 'team and fans,' which is plural, so 'are' is correct. The verb agrees with the plural noun." }, { "role...
[ { "prefix_pair_index": 1, "probes": [ { "id": "1505:coreference_disambiguation:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Can you explain the subject-verb agreement in 'The...
{ "source_dataset": "ChatAlpaca", "source_row_id": "1505", "scenario_id": "coreference_disambiguation", "scenario_title": "Coreference Disambiguation", "scenario_instruction": "Introduce at least two plausible referents of the same type, then make a later user follow-up rely on a pronoun or shorthand whose me...
827:hard_topic_reset:predictive
[ { "role": "user", "content": "Can you summarize the relationship between Vietnam and the global coffee trade?" }, { "role": "assistant", "content": "Vietnam is the world's second largest coffee producer after Brazil. They primarily grow Robusta beans, which contributes significantly to the count...
[ { "prefix_pair_index": 1, "probes": [ { "id": "827:hard_topic_reset:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Can you summarize the relationship between Vietnam and the gl...
{ "source_dataset": "ChatAlpaca", "source_row_id": "827", "scenario_id": "hard_topic_reset", "scenario_title": "Hard Topic Reset", "scenario_instruction": "Make the user abruptly switch to a different topic for one turn, then later return to the original topic. The assistant must not let the temporary switch ...
6393:two_thread_tracking:predictive
[ { "role": "user", "content": "Where do you think blockchain has the biggest impact in the real world?" }, { "role": "assistant", "content": "Logistics is a huge one. For example, tracking cargo from port to warehouse ensures no tampering with goods." }, { "role": "user", "content": "...
[ { "prefix_pair_index": 1, "probes": [ { "id": "6393:two_thread_tracking:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Where do you think blockchain has the biggest impact in t...
{ "source_dataset": "ChatAlpaca", "source_row_id": "6393", "scenario_id": "two_thread_tracking", "scenario_title": "Two Thread Tracking", "scenario_instruction": "Create two interleaved subproblems inside one conversation and ensure the assistant must keep both threads straight rather than collapsing them tog...
9091:self_correction_memory:predictive
[ { "role": "user", "content": "Insert a comma to make the following sentence grammatically correct: The bakery, that I visited yesterday, had freshly made croissants." }, { "role": "assistant", "content": "The bakery that I visited yesterday had freshly made croissants." }, { "role": "use...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9091:self_correction_memory:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Insert a comma to make the following sentence grammati...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9091", "scenario_id": "self_correction_memory", "scenario_title": "Self-correction Memory", "scenario_instruction": "Let the assistant revise or correct something earlier, then make a later turn depend on remembering the corrected version rather than the ori...
4263:long_explanation_then_specific_followup:predictive
[ { "role": "user", "content": "I'm drafting a poem about the night sky. Can you help me identify the core colors used to describe a stormy night versus a clear night?" }, { "role": "assistant", "content": "For a clear night, poets often use words like azure, violet, and deep blue to represent the...
[ { "prefix_pair_index": 1, "probes": [ { "id": "4263:long_explanation_then_specific_followup:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I'm drafting a poem about the night s...
{ "source_dataset": "ChatAlpaca", "source_row_id": "4263", "scenario_id": "long_explanation_then_specific_followup", "scenario_title": "Long Explanation Then Specific Follow-up", "scenario_instruction": "Make the assistant provide a fairly detailed middle-turn explanation, then have the user ask a specific fo...
9221:early_fact_late_recall:predictive
[ { "role": "user", "content": "Name 5 emotional states." }, { "role": "assistant", "content": "Joy, sadness, fear, anger, and disgust. Also, for the rest of this conversation, I will stick to work-related examples only." }, { "role": "user", "content": "Give me an example of joy in a ...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9221:early_fact_late_recall:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Name 5 emotional states.", "evidence": "Name 5...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9221", "scenario_id": "early_fact_late_recall", "scenario_title": "Early Fact, Late Recall", "scenario_instruction": "Introduce a concrete fact, label, preference, or commitment early, then make a later user turn depend on recalling it after at least one dis...
4815:late_constraint_reversal:predictive
[ { "role": "user", "content": "Explain the difference between critical thinking and analytical thinking." }, { "role": "assistant", "content": "Critical thinking involves evaluating assumptions and biases to form a judgment, while analytical thinking focuses on breaking down complex problems into...
[ { "prefix_pair_index": 1, "probes": [ { "id": "4815:late_constraint_reversal:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Explain the difference between critical thinking and...
{ "source_dataset": "ChatAlpaca", "source_row_id": "4815", "scenario_id": "late_constraint_reversal", "scenario_title": "Late Constraint Reversal", "scenario_instruction": "Make the assistant start solving one version of the problem, then add a late user constraint that forces the assistant to adapt without f...
1950:preserve_opening_retarget_later:predictive
[ { "role": "user", "content": "Compute the following expression: 10 + 6 * 7 - 58" }, { "role": "assistant", "content": "10 + 6 * 7 - 58 = 48" }, { "role": "user", "content": "Since I got 48, what would the total be if I added 100 to it?" }, { "role": "assistant", "content"...
[ { "prefix_pair_index": 1, "probes": [ { "id": "1950:preserve_opening_retarget_later:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Compute the following expression: 10 + 6 * 7 ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "1950", "scenario_id": "preserve_opening_retarget_later", "scenario_title": "Preserve Opening, Retarget Later", "scenario_instruction": "Keep the first user message exactly, but rewrite later turns so the conversation develops into a harder memory-dependent f...
9721:user_attention_test:predictive
[ { "role": "user", "content": "Construct an algorithm to solve the given problem: Problem: Calculate the sum of all multiples of 3 and 5 between 1 and 100" }, { "role": "assistant", "content": "Algorithm: 1. Initialize Sum as 0. 2. Set a variable i to 0. 3. While i is less than or equal to 100, d...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9721:user_attention_test:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Construct an algorithm to solve the given problem: Proble...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9721", "scenario_id": "user_attention_test", "scenario_title": "User Attention Test", "scenario_instruction": "Make the user explicitly test whether the assistant was paying attention to an earlier detail or instruction.", "must_have": [ "The user must...
1355:same_surface_final_question_new_context:predictive
[ { "role": "user", "content": "I've been considering switching my approach from a low-carb keto plan to something more sustainable." }, { "role": "assistant", "content": "That's a wise decision for long-term health. Have you looked into the intermittent fasting protocol?" }, { "role": "us...
[ { "prefix_pair_index": 1, "probes": [ { "id": "1355:same_surface_final_question_new_context:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "I've been considering switching my ap...
{ "source_dataset": "ChatAlpaca", "source_row_id": "1355", "scenario_id": "same_surface_final_question_new_context", "scenario_title": "Same Surface Final Question, New Context", "scenario_instruction": "Reshape the earlier history so the final user message keeps nearly the same surface form but now refers to...
9518:topic_pivot_then_return:predictive
[ { "role": "user", "content": "How is the backpropagation algorithm utilized specifically for training object detectors in autonomous driving systems?" }, { "role": "assistant", "content": "The backpropagation algorithm calculates the gradient of the loss function with respect to the model's weig...
[ { "prefix_pair_index": 1, "probes": [ { "id": "9518:topic_pivot_then_return:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "How is the backpropagation algorithm utilized specifi...
{ "source_dataset": "ChatAlpaca", "source_row_id": "9518", "scenario_id": "topic_pivot_then_return", "scenario_title": "Topic Pivot Then Return", "scenario_instruction": "Let the conversation briefly pivot to a secondary issue, then return to the original problem in a way that requires preserving the first th...
3023:late_constraint_reversal:predictive
[ { "role": "user", "content": "Classify the painting as either abstract or landscape: ![The-Ladies-of-the-Village-William-Henderson.jpg](attachment:The-Ladies-of-the-Village-William-Henderson.jpg)" }, { "role": "assistant", "content": "Abstract." }, { "role": "user", "content": "Can y...
[ { "prefix_pair_index": 1, "probes": [ { "id": "3023:late_constraint_reversal:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Classify the painting as either abstract or landscap...
{ "source_dataset": "ChatAlpaca", "source_row_id": "3023", "scenario_id": "late_constraint_reversal", "scenario_title": "Late Constraint Reversal", "scenario_instruction": "Make the assistant start solving one version of the problem, then add a late user constraint that forces the assistant to adapt without f...
1149:topic_pivot_then_return:predictive
[ { "role": "user", "content": "Explain the following business challenge: Restaurant industry faces numerous challenges related to rising labor costs, increasing competition and slow population growth." }, { "role": "assistant", "content": "Yes, those challenges threaten long-term viability. To su...
[ { "prefix_pair_index": 1, "probes": [ { "id": "1149:topic_pivot_then_return:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Explain the following business challenge: Restaurant ...
{ "source_dataset": "ChatAlpaca", "source_row_id": "1149", "scenario_id": "topic_pivot_then_return", "scenario_title": "Topic Pivot Then Return", "scenario_instruction": "Let the conversation briefly pivot to a secondary issue, then return to the original problem in a way that requires preserving the first th...
3828:long_explanation_then_specific_followup:predictive
[ { "role": "user", "content": "Given the following words, create a descriptive sentence: Rose, sweet, fragrance." }, { "role": "assistant", "content": "The sweet fragrance of the rose filled the air." }, { "role": "user", "content": "Can you suggest any other flowers that have a pleas...
[ { "prefix_pair_index": 1, "probes": [ { "id": "3828:long_explanation_then_specific_followup:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Given the following words, create a d...
{ "source_dataset": "ChatAlpaca", "source_row_id": "3828", "scenario_id": "long_explanation_then_specific_followup", "scenario_title": "Long Explanation Then Specific Follow-up", "scenario_instruction": "Make the assistant provide a fairly detailed middle-turn explanation, then have the user ask a specific fo...
854:user_attention_test:predictive
[ { "role": "user", "content": "Explain the symbolism in this piece of artwork: A painting with a mixture of green, yellow, and red." }, { "role": "assistant", "content": "The painting could represent growth, with green for new beginnings, yellow for joy, and red for passion. Culturally, green oft...
[ { "prefix_pair_index": 1, "probes": [ { "id": "854:user_attention_test:predictive:p1:first_user", "family": "first_user_request", "question": "What was the user's first request in this conversation?", "answer": "Explain the symbolism in this piece of artwork: A painting...
{ "source_dataset": "ChatAlpaca", "source_row_id": "854", "scenario_id": "user_attention_test", "scenario_title": "User Attention Test", "scenario_instruction": "Make the user explicitly test whether the assistant was paying attention to an earlier detail or instruction.", "must_have": [ "The user must ...