Dataset Viewer
Auto-converted to Parquet Duplicate
instruction
stringlengths
16
91
answer
stringclasses
36 values
tag
stringclasses
35 values
split_kind
stringclasses
14 values
text
stringlengths
53
134
Give the code/prose label: This sentence is plain English.
prose
classify_prose
task_seen_train
Instruction: Give the code/prose label: This sentence is plain English. Answer: prose <END>
Can water become frozen? Give a reason.
Yes, because water can become ice when it is cold enough.
yes_water_freeze
train_seen
Instruction: Can water become frozen? Give a reason. Answer: Yes, because water can become ice when it is cold enough. <END>
The prompt gives this password: black. Question: Which password does it give?
black.
context_password
context_upstream_regression_train
Instruction: The prompt gives this password: black. Question: Which password does it give? Answer: black. <END>
Context: name = Kai; password = gold. Question: What is the password?
gold.
context_password
context_seen_train
Instruction: Context: name = Kai; password = gold. Question: What is the password? Answer: gold. <END>
User: My name is Jordan. User: What is my name?
Jordan.
context_name
context_seen_train
Instruction: User: My name is Jordan. User: What is my name? Answer: Jordan. <END>
Can water freeze? Answer yes or no with a reason.
Yes, because water can become ice when it is cold enough.
yes_water_freeze
train_seen
Instruction: Can water freeze? Answer yes or no with a reason. Answer: Yes, because water can become ice when it is cold enough. <END>
Choose the sentiment label for: I enjoy this.
positive
sentiment_positive
task_seen_train
Instruction: Choose the sentiment label for: I enjoy this. Answer: positive <END>
Name what Evan ate today.
I don't know from the given information.
idk_food_today
task_upstream_regression_train
Instruction: Name what Evan ate today. Answer: I don't know from the given information. <END>
Give the sentiment label only: This is awful.
negative
sentiment_negative
task_upstream_regression_train
Instruction: Give the sentiment label only: This is awful. Answer: negative <END>
Context line says password -> silver. Question: What password value follows the arrow?
silver.
context_password
context_seen_train
Instruction: Context line says password -> silver. Question: What password value follows the arrow? Answer: silver. <END>
The password supplied in this prompt is purple. Question: Which password was supplied?
purple.
context_password
context_seen_train
Instruction: The password supplied in this prompt is purple. Question: Which password was supplied? Answer: purple. <END>
Choose the sentiment label for: This is bad.
negative
sentiment_negative
task_seen_train
Instruction: Choose the sentiment label for: This is bad. Answer: negative <END>
The prompt gives this password: red. Question: Which password does it give?
red.
context_password
context_upstream_regression_train
Instruction: The prompt gives this password: red. Question: Which password does it give? Answer: red. <END>
Provided name: Evan. Question: What name was provided?
Evan.
context_name
context_upstream_regression_train
Instruction: Provided name: Evan. Question: What name was provided? Answer: Evan. <END>
Facts: name is Jordan. password is green. Question: Which password was given?
green.
context_password
context_seen_train
Instruction: Facts: name is Jordan. password is green. Question: Which password was given? Answer: green. <END>
Explain a return statement in Python.
A return statement sends a value back from a function.
return_statement
train_seen
Instruction: Explain a return statement in Python. Answer: A return statement sends a value back from a function. <END>
What type of text is this: The file is on the desk.
prose
classify_prose
task_seen_train
Instruction: What type of text is this: The file is on the desk. Answer: prose <END>
User: Call me Maya. User: What should you call me?
Maya.
context_name
context_seen_train
Instruction: User: Call me Maya. User: What should you call me? Answer: Maya. <END>
What type of text is this: print('hello')
code
classify_code
task_seen_train
Instruction: What type of text is this: print('hello') Answer: code <END>
Positive or negative sentiment: This helped a lot.
positive
sentiment_positive
task_seen_train
Instruction: Positive or negative sentiment: This helped a lot. Answer: positive <END>
What type of text is this: x = add(1, 2)
code
classify_code
task_seen_train
Instruction: What type of text is this: x = add(1, 2) Answer: code <END>
Return the product using multiply(a, b).
def multiply(a, b): return a * b
code_multiply
train_seen
Instruction: Return the product using multiply(a, b). Answer: def multiply(a, b): return a * b <END>
Classify sentiment: That was useful.
positive
sentiment_positive
task_seen_train
Instruction: Classify sentiment: That was useful. Answer: positive <END>
Conversation: User: I am Evan. User: Who am I?
Evan.
context_name
context_seen_train
Instruction: Conversation: User: I am Evan. User: Who am I? Answer: Evan. <END>
Classify this as positive or negative: The result is terrible.
negative
sentiment_negative
task_seen_train
Instruction: Classify this as positive or negative: The result is terrible. Answer: negative <END>
Classify this as code or prose: Please read the current file first.
prose
classify_prose
task_seen_train
Instruction: Classify this as code or prose: Please read the current file first. Answer: prose <END>
User: The password is orange. User: What is the password?
orange.
context_password
context_seen_train
Instruction: User: The password is orange. User: What is the password? Answer: orange. <END>
Answer only positive or negative: This helped a lot.
positive
sentiment_positive
task_seen_train
Instruction: Answer only positive or negative: This helped a lot. Answer: positive <END>
Context: name = Sam; password = green. Question: What is the name?
Sam.
context_name
context_seen_train
Instruction: Context: name = Sam; password = green. Question: What is the name? Answer: Sam. <END>
Classify sentiment: This is broken.
negative
sentiment_negative
task_upstream_regression_train
Instruction: Classify sentiment: This is broken. Answer: negative <END>
Which city is France's capital?
Paris.
capital_france
train_seen
Instruction: Which city is France's capital? Answer: Paris. <END>
Classify this as positive or negative: This is broken.
negative
sentiment_negative
task_upstream_regression_train
Instruction: Classify this as positive or negative: This is broken. Answer: negative <END>
User: Use the name Nina. User: What name did I give?
Nina.
context_name
context_seen_train
Instruction: User: Use the name Nina. User: What name did I give? Answer: Nina. <END>
Context: password = orange. Question: What is the password?
orange.
context_password
context_seen_train
Instruction: Context: password = orange. Question: What is the password? Answer: orange. <END>
Positive or negative sentiment: This is broken.
negative
sentiment_negative
task_seen_train
Instruction: Positive or negative sentiment: This is broken. Answer: negative <END>
What sentiment is this: The result is excellent.
positive
sentiment_positive
task_seen_train
Instruction: What sentiment is this: The result is excellent. Answer: positive <END>
Formal rewrite: I can't make it.
I am unable to attend.
rewrite_professional_attend
task_upstream_regression_train
Instruction: Formal rewrite: I can't make it. Answer: I am unable to attend. <END>
What type of text is this: return value
code
classify_code
task_seen_train
Instruction: What type of text is this: return value Answer: code <END>
If you do not have enough current evidence, what is the right move?
Ask for the missing evidence.
evidence_missing
train_seen
Instruction: If you do not have enough current evidence, what is the right move? Answer: Ask for the missing evidence. <END>
Remember inside this prompt: call the user Evan. Question: What should you call the user?
Evan.
context_name
context_seen_train
Instruction: Remember inside this prompt: call the user Evan. Question: What should you call the user? Answer: Evan. <END>
Classify input: def add(a, b): return a + b
code
classify_code
task_seen_train
Instruction: Classify input: def add(a, b): return a + b Answer: code <END>
Context: password = black. Question: What is the password?
black.
context_password
context_seen_train
Instruction: Context: password = black. Question: What is the password? Answer: black. <END>
Conversation: User: Use green as the password. User: What password did I give?
green.
context_password
context_seen_train
Instruction: Conversation: User: Use green as the password. User: What password did I give? Answer: green. <END>
Line one says PASSWORD=red. Question: What is PASSWORD?
red.
context_password
context_upstream_regression_train
Instruction: Line one says PASSWORD=red. Question: What is PASSWORD? Answer: red. <END>
What was Evan's meal today?
I don't know from the given information.
idk_food_today
task_upstream_regression_train
Instruction: What was Evan's meal today? Answer: I don't know from the given information. <END>
Before changing code, what should happen first?
Inspect the current files first.
inspect_first
train_seen
Instruction: Before changing code, what should happen first? Answer: Inspect the current files first. <END>
Classify sentiment: This is strong.
positive
sentiment_positive
task_seen_train
Instruction: Classify sentiment: This is strong. Answer: positive <END>
The supplied name is Kai. Question: What is the supplied name?
Kai.
context_name
context_seen_train
Instruction: The supplied name is Kai. Question: What is the supplied name? Answer: Kai. <END>
Make a Python function named subtract that subtracts b from a.
def subtract(a, b): return a - b
code_subtract
train_seen
Instruction: Make a Python function named subtract that subtracts b from a. Answer: def subtract(a, b): return a - b <END>
Classify sentiment: This is broken.
negative
sentiment_negative
task_seen_train
Instruction: Classify sentiment: This is broken. Answer: negative <END>
Is this sentiment positive or negative: This feels wrong.
negative
sentiment_negative
task_seen_train
Instruction: Is this sentiment positive or negative: This feels wrong. Answer: negative <END>
The name supplied in this prompt is Kai. Question: Which name was supplied?
Kai.
context_name
context_upstream_regression_train
Instruction: The name supplied in this prompt is Kai. Question: Which name was supplied? Answer: Kai. <END>
Make a Python function named divide that divides a by b.
def divide(a, b): return a / b
code_divide
train_seen
Instruction: Make a Python function named divide that divides a by b. Answer: def divide(a, b): return a / b <END>
Give only the type label for this: A clear sky is usually blue.
prose
classify_prose
task_seen_train
Instruction: Give only the type label for this: A clear sky is usually blue. Answer: prose <END>
Label the input type: return value
code
classify_code
task_seen_train
Instruction: Label the input type: return value Answer: code <END>
Polish this professionally: I won't be able to make it.
I am unable to attend.
rewrite_professional_attend
task_seen_train
Instruction: Polish this professionally: I won't be able to make it. Answer: I am unable to attend. <END>
Give the sentiment label only: This failed badly.
negative
sentiment_negative
task_seen_train
Instruction: Give the sentiment label only: This failed badly. Answer: negative <END>
Classify input: The user asked a direct question.
prose
classify_prose
task_seen_train
Instruction: Classify input: The user asked a direct question. Answer: prose <END>
How should you answer the user's request?
I answer the user's request directly.
assistant_role
tt639g_lite_v2_failure_repair_train
Instruction: How should you answer the user's request? Answer: I answer the user's request directly. <END>
Type label: Please read the current file first.
prose
classify_prose
task_seen_train
Instruction: Type label: Please read the current file first. Answer: prose <END>
Remember password value orange. Question: Which password value was stated?
orange.
context_password
context_seen_train
Instruction: Remember password value orange. Question: Which password value was stated? Answer: orange. <END>
What sentiment is this: This is broken.
negative
sentiment_negative
task_upstream_regression_train
Instruction: What sentiment is this: This is broken. Answer: negative <END>
Facts: password is blue. name is Nina. Question: Which name was given?
Nina.
context_name
context_seen_train
Instruction: Facts: password is blue. name is Nina. Question: Which name was given? Answer: Nina. <END>
Context: name = Riley; password = red. Question: What is the name?
Riley.
context_name
context_seen_train
Instruction: Context: name = Riley; password = red. Question: What is the name? Answer: Riley. <END>
Answer only positive or negative: The fix worked.
positive
sentiment_positive
task_seen_train
Instruction: Answer only positive or negative: The fix worked. Answer: positive <END>
Return the product using multiply(a, b).
def multiply(a, b): return a * b
code_multiply
train_seen
Instruction: Return the product using multiply(a, b). Answer: def multiply(a, b): return a * b <END>
What sentiment is this: I really like this.
positive
sentiment_positive
task_upstream_regression_train
Instruction: What sentiment is this: I really like this. Answer: positive <END>
How should you help?
I answer the user's request directly.
assistant_role
train_seen
Instruction: How should you help? Answer: I answer the user's request directly. <END>
This prompt supplies the password purple. Question: Which password is supplied?
purple.
context_password
context_seen_train
Instruction: This prompt supplies the password purple. Question: Which password is supplied? Answer: purple. <END>
Give only the type label for this: The appointment was moved to Friday.
prose
classify_prose
task_upstream_regression_train
Instruction: Give only the type label for this: The appointment was moved to Friday. Answer: prose <END>
What sentiment is this: I enjoy this.
positive
sentiment_positive
task_seen_train
Instruction: What sentiment is this: I enjoy this. Answer: positive <END>
Classify input: The meeting starts tomorrow.
prose
classify_prose
task_seen_train
Instruction: Classify input: The meeting starts tomorrow. Answer: prose <END>
Give the code/prose label: print('hello')
code
classify_code
task_seen_train
Instruction: Give the code/prose label: print('hello') Answer: code <END>
Context: name = Maya; password = silver. Question: What is the password?
silver.
context_password
context_seen_train
Instruction: Context: name = Maya; password = silver. Question: What is the password? Answer: silver. <END>
Make this concise: I am going to arrive late today.
I will be late today.
rewrite_short_late
task_upstream_regression_train
Instruction: Make this concise: I am going to arrive late today. Answer: I will be late today. <END>
Facts: name is Sam. password is blue. Question: Which password was given?
blue.
context_password
context_seen_train
Instruction: Facts: name is Sam. password is blue. Question: Which password was given? Answer: blue. <END>
How should you help?
I answer the user's request directly.
assistant_role
tt641_v2_guard_train
Instruction: How should you help? Answer: I answer the user's request directly. <END>
Sentiment of this text: The result is terrible.
negative
sentiment_negative
task_seen_train
Instruction: Sentiment of this text: The result is terrible. Answer: negative <END>
Sentiment label only: This feels solid.
positive
sentiment_positive
task_seen_train
Instruction: Sentiment label only: This feels solid. Answer: positive <END>
This prompt supplies the password red. Question: Which password is supplied?
red.
context_password
context_seen_train
Instruction: This prompt supplies the password red. Question: Which password is supplied? Answer: red. <END>
The password supplied in this prompt is green. Question: Which password was supplied?
green.
context_password
context_upstream_regression_train
Instruction: The password supplied in this prompt is green. Question: Which password was supplied? Answer: green. <END>
What type of text is this: return value
code
classify_code
task_seen_train
Instruction: What type of text is this: return value Answer: code <END>
Type label: def add(a, b): return a + b
code
classify_code
task_seen_train
Instruction: Type label: def add(a, b): return a + b Answer: code <END>
Choose one label, code or prose: for i in range(3): print(i)
code
classify_code
task_seen_train
Instruction: Choose one label, code or prose: for i in range(3): print(i) Answer: code <END>
Facts: password is purple. name is Maya. Question: Which name was given?
Maya.
context_name
context_seen_train
Instruction: Facts: password is purple. name is Maya. Question: Which name was given? Answer: Maya. <END>
Facts: password is gold. name is Jordan. Question: Which name was given?
Jordan.
context_name
context_seen_train
Instruction: Facts: password is gold. name is Jordan. Question: Which name was given? Answer: Jordan. <END>
Make this concise: I am going to arrive late today.
I will be late today.
rewrite_short_late
task_upstream_regression_train
Instruction: Make this concise: I am going to arrive late today. Answer: I will be late today. <END>
Formal version: I cannot make it.
I am unable to attend.
rewrite_professional_attend
task_upstream_regression_train
Instruction: Formal version: I cannot make it. Answer: I am unable to attend. <END>
What sentiment is this: That was helpful.
positive
sentiment_positive
task_seen_train
Instruction: What sentiment is this: That was helpful. Answer: positive <END>
Give the sentiment label only: This is bad.
negative
sentiment_negative
task_seen_train
Instruction: Give the sentiment label only: This is bad. Answer: negative <END>
Write the ADD function.
def add(a, b): return a + b
code_add
train_seen
Instruction: Write the ADD function. Answer: def add(a, b): return a + b <END>
What lunch did Evan say he ate?
I don't know from the given information.
idk_food_today
task_seen_train
Instruction: What lunch did Evan say he ate? Answer: I don't know from the given information. <END>
Facts: name is Nina. password is green. Question: Which password was given?
green.
context_password
context_seen_train
Instruction: Facts: name is Nina. password is green. Question: Which password was given? Answer: green. <END>
Sentiment label only: This is broken.
negative
sentiment_negative
task_seen_train
Instruction: Sentiment label only: This is broken. Answer: negative <END>
Label the sentiment: This is good.
positive
sentiment_positive
task_seen_train
Instruction: Label the sentiment: This is good. Answer: positive <END>
Rewrite professionally: I can't make it.
I am unable to attend.
rewrite_professional_attend
task_upstream_regression_train
Instruction: Rewrite professionally: I can't make it. Answer: I am unable to attend. <END>
Facts: password is orange. name is Sam. Question: Which name was given?
Sam.
context_name
context_seen_train
Instruction: Facts: password is orange. name is Sam. Question: Which name was given? Answer: Sam. <END>
Positive or negative sentiment: That failed again.
negative
sentiment_negative
task_upstream_regression_train
Instruction: Positive or negative sentiment: That failed again. Answer: negative <END>
What sentiment is this: This works great.
positive
sentiment_positive
task_seen_train
Instruction: What sentiment is this: This works great. Answer: positive <END>
Give only the type label for this: The appointment was moved to Friday.
prose
classify_prose
task_upstream_regression_train
Instruction: Give only the type label for this: The appointment was moved to Friday. Answer: prose <END>
End of preview. Expand in Data Studio

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

TT639G Recombined Tiny Assistant v1

Recombines isolated proof rungs:

  • TT638D code behavior + dyadic/Mercy proof upstream
  • TT639E2 context-copy behavior
  • TT639F3 task-routing behavior
  • simple rule/Q&A behavior

Blocking dense gates:

  • seen_combined_pass
  • upstream_regression_pass
  • mixed_heldout_pass
  • anti_collision_pass

Do not run dyadic/Mercy compare unless all four gates pass.

Downloads last month
57