source stringclasses 18
values | messages listlengths 2 2 | rubric stringlengths 89 2.21k | score float64 0 1 | weighted_score float64 0 1 |
|---|---|---|---|---|
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must provide a single word as the fill-in answer for the crossword puzzle. [Hard Rule]
2. The response must select a word that conveys a strong sense of determination toward one’s desired profession. [Hard Rule]
3. The response must not propose a word that is a synonym for “thrill” or “enthusiasm.” [Hard Rule]
4. The response should present the answer in a clear, easily identifiable format. [Principle]
5. The response should include a direct justification linking the chosen word to the clue’s definition. [Principle]
6. The response should be concise and avoid extraneous or off-topic commentary. [Principle]
7. The response should maintain a professional, task-focused tone without unnecessary conversational fillers. [Principle] | 0.714286 | 0.565217 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must provide a single word as the fill-in answer for the crossword puzzle. [Hard Rule]
2. The response must select a word that conveys a strong sense of determination toward one’s desired profession. [Hard Rule]
3. The response must not propose a word that is a synonym for “thrill” or “enthusiasm.” [Hard Rule]
4. The response should present the answer in a clear, easily identifiable format. [Principle]
5. The response should include a direct justification linking the chosen word to the clue’s definition. [Principle]
6. The response should be concise and avoid extraneous or off-topic commentary. [Principle]
7. The response should maintain a professional, task-focused tone without unnecessary conversational fillers. [Principle] | 0.428571 | 0.4 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must present the answer in a two-column table with headers “Incorrect Sentence” and “Corrected Sentence” matching the requested wording and format. [Hard Rule]
2. The response must include the original sentence under the “Incorrect Sentence” column and its corrected version under the “Corrected Sentence” column. [Hard Rule]
3. The response must use proper English grammar in the corrected sentence. [Hard Rule]
4. The response should remain focused on the user’s request without introducing unrelated explanations or commentary. [Principle]
5. The response should organize information clearly and in a straightforward structure to facilitate quick comprehension. [Principle]
6. The response should use labels and terminology consistently and precisely as specified by the user. [Principle]
7. The response should ensure that all provided content is accurate and free of errors. [Principle] | 0.428571 | 0.4 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must present the answer in a two-column table with headers “Incorrect Sentence” and “Corrected Sentence” matching the requested wording and format. [Hard Rule]
2. The response must include the original sentence under the “Incorrect Sentence” column and its corrected version under the “Corrected Sentence” column. [Hard Rule]
3. The response must use proper English grammar in the corrected sentence. [Hard Rule]
4. The response should remain focused on the user’s request without introducing unrelated explanations or commentary. [Principle]
5. The response should organize information clearly and in a straightforward structure to facilitate quick comprehension. [Principle]
6. The response should use labels and terminology consistently and precisely as specified by the user. [Principle]
7. The response should ensure that all provided content is accurate and free of errors. [Principle] | 0.428571 | 0.473684 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must present the answer in a two-column table with headers “Incorrect Sentence” and “Corrected Sentence” matching the requested wording and format. [Hard Rule]
2. The response must include the original sentence under the “Incorrect Sentence” column and its corrected version under the “Corrected Sentence” column. [Hard Rule]
3. The response must use proper English grammar in the corrected sentence. [Hard Rule]
4. The response should remain focused on the user’s request without introducing unrelated explanations or commentary. [Principle]
5. The response should organize information clearly and in a straightforward structure to facilitate quick comprehension. [Principle]
6. The response should use labels and terminology consistently and precisely as specified by the user. [Principle]
7. The response should ensure that all provided content is accurate and free of errors. [Principle] | 0.428571 | 0.4 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must present the answer in a two-column table with headers “Incorrect Sentence” and “Corrected Sentence” matching the requested wording and format. [Hard Rule]
2. The response must include the original sentence under the “Incorrect Sentence” column and its corrected version under the “Corrected Sentence” column. [Hard Rule]
3. The response must use proper English grammar in the corrected sentence. [Hard Rule]
4. The response should remain focused on the user’s request without introducing unrelated explanations or commentary. [Principle]
5. The response should organize information clearly and in a straightforward structure to facilitate quick comprehension. [Principle]
6. The response should use labels and terminology consistently and precisely as specified by the user. [Principle]
7. The response should ensure that all provided content is accurate and free of errors. [Principle] | 1 | 1 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response proposes advanced machine learning techniques to enhance real-time object recognition and path planning. [Hard Rule]
2. The response proposes incorporating force feedback sensors that adapt grip strength based on object weight and fragility. [Hard Rule]
3. The response explores integrating a feedback control system to monitor and adjust the robotic arm’s position and orientation for improved precision and accuracy. [Hard Rule]
4. The response demonstrates completeness by addressing each aspect of the task thoroughly. [Principle]
5. The response employs clear and logical organization to present information in a structured format. [Principle]
6. The response includes concrete and relevant methodological examples to support proposed solutions. [Principle]
7. The response integrates multiple system components cohesively to illustrate how they interact. [Principle]
8. The response incorporates mechanisms for adaptation or continuous improvement over time. [Principle]
9. The response justifies the selection of techniques by explaining their relevance and contribution to the overall objectives. [Principle]
10. The response considers the fusion or combined analysis of different sensor or data sources to enhance system performance. [Principle] | 0.3 | 0.405405 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response proposes advanced machine learning techniques to enhance real-time object recognition and path planning. [Hard Rule]
2. The response proposes incorporating force feedback sensors that adapt grip strength based on object weight and fragility. [Hard Rule]
3. The response explores integrating a feedback control system to monitor and adjust the robotic arm’s position and orientation for improved precision and accuracy. [Hard Rule]
4. The response demonstrates completeness by addressing each aspect of the task thoroughly. [Principle]
5. The response employs clear and logical organization to present information in a structured format. [Principle]
6. The response includes concrete and relevant methodological examples to support proposed solutions. [Principle]
7. The response integrates multiple system components cohesively to illustrate how they interact. [Principle]
8. The response incorporates mechanisms for adaptation or continuous improvement over time. [Principle]
9. The response justifies the selection of techniques by explaining their relevance and contribution to the overall objectives. [Principle]
10. The response considers the fusion or combined analysis of different sensor or data sources to enhance system performance. [Principle] | 0.9 | 0.91358 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must present the answer in a two-column table with headers “Incorrect Sentence” and “Corrected Sentence” matching the requested wording and format. [Hard Rule]
2. The response must include the original sentence under the “Incorrect Sentence” column and its corrected version under the “Corrected Sentence” column. [Hard Rule]
3. The response must use proper English grammar in the corrected sentence. [Hard Rule]
4. The response should remain focused on the user’s request without introducing unrelated explanations or commentary. [Principle]
5. The response should organize information clearly and in a straightforward structure to facilitate quick comprehension. [Principle]
6. The response should use labels and terminology consistently and precisely as specified by the user. [Principle]
7. The response should ensure that all provided content is accurate and free of errors. [Principle] | 0.857143 | 0.87931 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must present the answer in a two-column table with headers “Incorrect Sentence” and “Corrected Sentence” matching the requested wording and format. [Hard Rule]
2. The response must include the original sentence under the “Incorrect Sentence” column and its corrected version under the “Corrected Sentence” column. [Hard Rule]
3. The response must use proper English grammar in the corrected sentence. [Hard Rule]
4. The response should remain focused on the user’s request without introducing unrelated explanations or commentary. [Principle]
5. The response should organize information clearly and in a straightforward structure to facilitate quick comprehension. [Principle]
6. The response should use labels and terminology consistently and precisely as specified by the user. [Principle]
7. The response should ensure that all provided content is accurate and free of errors. [Principle] | 0.285714 | 0.315789 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response proposes advanced machine learning techniques to enhance real-time object recognition and path planning. [Hard Rule]
2. The response proposes incorporating force feedback sensors that adapt grip strength based on object weight and fragility. [Hard Rule]
3. The response explores integrating a feedback control system to monitor and adjust the robotic arm’s position and orientation for improved precision and accuracy. [Hard Rule]
4. The response demonstrates completeness by addressing each aspect of the task thoroughly. [Principle]
5. The response employs clear and logical organization to present information in a structured format. [Principle]
6. The response includes concrete and relevant methodological examples to support proposed solutions. [Principle]
7. The response integrates multiple system components cohesively to illustrate how they interact. [Principle]
8. The response incorporates mechanisms for adaptation or continuous improvement over time. [Principle]
9. The response justifies the selection of techniques by explaining their relevance and contribution to the overall objectives. [Principle]
10. The response considers the fusion or combined analysis of different sensor or data sources to enhance system performance. [Principle] | 0.4 | 0.4625 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response proposes advanced machine learning techniques to enhance real-time object recognition and path planning. [Hard Rule]
2. The response proposes incorporating force feedback sensors that adapt grip strength based on object weight and fragility. [Hard Rule]
3. The response explores integrating a feedback control system to monitor and adjust the robotic arm’s position and orientation for improved precision and accuracy. [Hard Rule]
4. The response demonstrates completeness by addressing each aspect of the task thoroughly. [Principle]
5. The response employs clear and logical organization to present information in a structured format. [Principle]
6. The response includes concrete and relevant methodological examples to support proposed solutions. [Principle]
7. The response integrates multiple system components cohesively to illustrate how they interact. [Principle]
8. The response incorporates mechanisms for adaptation or continuous improvement over time. [Principle]
9. The response justifies the selection of techniques by explaining their relevance and contribution to the overall objectives. [Principle]
10. The response considers the fusion or combined analysis of different sensor or data sources to enhance system performance. [Principle] | 1 | 1 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response proposes advanced machine learning techniques to enhance real-time object recognition and path planning. [Hard Rule]
2. The response proposes incorporating force feedback sensors that adapt grip strength based on object weight and fragility. [Hard Rule]
3. The response explores integrating a feedback control system to monitor and adjust the robotic arm’s position and orientation for improved precision and accuracy. [Hard Rule]
4. The response demonstrates completeness by addressing each aspect of the task thoroughly. [Principle]
5. The response employs clear and logical organization to present information in a structured format. [Principle]
6. The response includes concrete and relevant methodological examples to support proposed solutions. [Principle]
7. The response integrates multiple system components cohesively to illustrate how they interact. [Principle]
8. The response incorporates mechanisms for adaptation or continuous improvement over time. [Principle]
9. The response justifies the selection of techniques by explaining their relevance and contribution to the overall objectives. [Principle]
10. The response considers the fusion or combined analysis of different sensor or data sources to enhance system performance. [Principle] | 0.4 | 0.493151 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response proposes advanced machine learning techniques to enhance real-time object recognition and path planning. [Hard Rule]
2. The response proposes incorporating force feedback sensors that adapt grip strength based on object weight and fragility. [Hard Rule]
3. The response explores integrating a feedback control system to monitor and adjust the robotic arm’s position and orientation for improved precision and accuracy. [Hard Rule]
4. The response demonstrates completeness by addressing each aspect of the task thoroughly. [Principle]
5. The response employs clear and logical organization to present information in a structured format. [Principle]
6. The response includes concrete and relevant methodological examples to support proposed solutions. [Principle]
7. The response integrates multiple system components cohesively to illustrate how they interact. [Principle]
8. The response incorporates mechanisms for adaptation or continuous improvement over time. [Principle]
9. The response justifies the selection of techniques by explaining their relevance and contribution to the overall objectives. [Principle]
10. The response considers the fusion or combined analysis of different sensor or data sources to enhance system performance. [Principle] | 1 | 1 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response proposes advanced machine learning techniques to enhance real-time object recognition and path planning. [Hard Rule]
2. The response proposes incorporating force feedback sensors that adapt grip strength based on object weight and fragility. [Hard Rule]
3. The response explores integrating a feedback control system to monitor and adjust the robotic arm’s position and orientation for improved precision and accuracy. [Hard Rule]
4. The response demonstrates completeness by addressing each aspect of the task thoroughly. [Principle]
5. The response employs clear and logical organization to present information in a structured format. [Principle]
6. The response includes concrete and relevant methodological examples to support proposed solutions. [Principle]
7. The response integrates multiple system components cohesively to illustrate how they interact. [Principle]
8. The response incorporates mechanisms for adaptation or continuous improvement over time. [Principle]
9. The response justifies the selection of techniques by explaining their relevance and contribution to the overall objectives. [Principle]
10. The response considers the fusion or combined analysis of different sensor or data sources to enhance system performance. [Principle] | 0.3 | 0.405405 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response proposes advanced machine learning techniques to enhance real-time object recognition and path planning. [Hard Rule]
2. The response proposes incorporating force feedback sensors that adapt grip strength based on object weight and fragility. [Hard Rule]
3. The response explores integrating a feedback control system to monitor and adjust the robotic arm’s position and orientation for improved precision and accuracy. [Hard Rule]
4. The response demonstrates completeness by addressing each aspect of the task thoroughly. [Principle]
5. The response employs clear and logical organization to present information in a structured format. [Principle]
6. The response includes concrete and relevant methodological examples to support proposed solutions. [Principle]
7. The response integrates multiple system components cohesively to illustrate how they interact. [Principle]
8. The response incorporates mechanisms for adaptation or continuous improvement over time. [Principle]
9. The response justifies the selection of techniques by explaining their relevance and contribution to the overall objectives. [Principle]
10. The response considers the fusion or combined analysis of different sensor or data sources to enhance system performance. [Principle] | 1 | 1 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must provide a single word as the fill-in answer for the crossword puzzle. [Hard Rule]
2. The response must select a word that conveys a strong sense of determination toward one’s desired profession. [Hard Rule]
3. The response must not propose a word that is a synonym for “thrill” or “enthusiasm.” [Hard Rule]
4. The response should present the answer in a clear, easily identifiable format. [Principle]
5. The response should include a direct justification linking the chosen word to the clue’s definition. [Principle]
6. The response should be concise and avoid extraneous or off-topic commentary. [Principle]
7. The response should maintain a professional, task-focused tone without unnecessary conversational fillers. [Principle] | 0.857143 | 0.8 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must provide a single word as the fill-in answer for the crossword puzzle. [Hard Rule]
2. The response must select a word that conveys a strong sense of determination toward one’s desired profession. [Hard Rule]
3. The response must not propose a word that is a synonym for “thrill” or “enthusiasm.” [Hard Rule]
4. The response should present the answer in a clear, easily identifiable format. [Principle]
5. The response should include a direct justification linking the chosen word to the clue’s definition. [Principle]
6. The response should be concise and avoid extraneous or off-topic commentary. [Principle]
7. The response should maintain a professional, task-focused tone without unnecessary conversational fillers. [Principle] | 0.714286 | 0.673913 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must correctly identify the figure of speech as hyperbole and explain how the given statement exemplifies intentional exaggeration. [Hard Rule]
2. The response clearly defines the relevant rhetorical device as a form of deliberate exaggeration used for effect. [Principle]
3. The response connects the definition to the example by illustrating how the statement’s content exceeds realistic bounds. [Principle]
4. The response explains the purpose or effect of the rhetorical device in context, such as emphasizing a point, creating vivid imagery, or adding humor. [Principle]
5. The response remains focused on analyzing the figure of speech without introducing unrelated commentary or advice. [Principle]
6. The response employs clear, concise language suitable for explanatory writing, avoiding unnecessary repetition or verbosity. [Principle]
7. The response maintains an objective and impersonal tone, avoiding direct audience addressing or normative judgments. [Principle] | 1 | 1 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must correctly identify the figure of speech as hyperbole and explain how the given statement exemplifies intentional exaggeration. [Hard Rule]
2. The response clearly defines the relevant rhetorical device as a form of deliberate exaggeration used for effect. [Principle]
3. The response connects the definition to the example by illustrating how the statement’s content exceeds realistic bounds. [Principle]
4. The response explains the purpose or effect of the rhetorical device in context, such as emphasizing a point, creating vivid imagery, or adding humor. [Principle]
5. The response remains focused on analyzing the figure of speech without introducing unrelated commentary or advice. [Principle]
6. The response employs clear, concise language suitable for explanatory writing, avoiding unnecessary repetition or verbosity. [Principle]
7. The response maintains an objective and impersonal tone, avoiding direct audience addressing or normative judgments. [Principle] | 0 | 0 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must correctly identify the figure of speech as hyperbole and explain how the given statement exemplifies intentional exaggeration. [Hard Rule]
2. The response clearly defines the relevant rhetorical device as a form of deliberate exaggeration used for effect. [Principle]
3. The response connects the definition to the example by illustrating how the statement’s content exceeds realistic bounds. [Principle]
4. The response explains the purpose or effect of the rhetorical device in context, such as emphasizing a point, creating vivid imagery, or adding humor. [Principle]
5. The response remains focused on analyzing the figure of speech without introducing unrelated commentary or advice. [Principle]
6. The response employs clear, concise language suitable for explanatory writing, avoiding unnecessary repetition or verbosity. [Principle]
7. The response maintains an objective and impersonal tone, avoiding direct audience addressing or normative judgments. [Principle] | 1 | 1 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must correctly identify the figure of speech as hyperbole and explain how the given statement exemplifies intentional exaggeration. [Hard Rule]
2. The response clearly defines the relevant rhetorical device as a form of deliberate exaggeration used for effect. [Principle]
3. The response connects the definition to the example by illustrating how the statement’s content exceeds realistic bounds. [Principle]
4. The response explains the purpose or effect of the rhetorical device in context, such as emphasizing a point, creating vivid imagery, or adding humor. [Principle]
5. The response remains focused on analyzing the figure of speech without introducing unrelated commentary or advice. [Principle]
6. The response employs clear, concise language suitable for explanatory writing, avoiding unnecessary repetition or verbosity. [Principle]
7. The response maintains an objective and impersonal tone, avoiding direct audience addressing or normative judgments. [Principle] | 0.571429 | 0.604167 |
ultrafeedback_evol_instruct | [
{
"content": "You are an impartial expert evaluator.\n\nYour task is to evaluate an assistant response using a given rubric.\n\n# Inputs\nYou will receive:\n- A user request\n- A single assistant response\n- A rubric containing evaluation criteria\n\n# Instructions\nFor EACH rubric criterion:\n1. Determine whet... | 1. The response must correctly identify the figure of speech as hyperbole and explain how the given statement exemplifies intentional exaggeration. [Hard Rule]
2. The response clearly defines the relevant rhetorical device as a form of deliberate exaggeration used for effect. [Principle]
3. The response connects the definition to the example by illustrating how the statement’s content exceeds realistic bounds. [Principle]
4. The response explains the purpose or effect of the rhetorical device in context, such as emphasizing a point, creating vivid imagery, or adding humor. [Principle]
5. The response remains focused on analyzing the figure of speech without introducing unrelated commentary or advice. [Principle]
6. The response employs clear, concise language suitable for explanatory writing, avoiding unnecessary repetition or verbosity. [Principle]
7. The response maintains an objective and impersonal tone, avoiding direct audience addressing or normative judgments. [Principle] | 0.285714 | 0.361702 |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 34