narrative string | question string | choices string | answer_index int64 | answer_choice string | task_source string | formatted_prompt list | response list | formatted_prompt__metadata list | response__metadata list | eval_result list | eval_result__metadata list | _judge_prompt__metadata list | _judge_response__metadata list | strategy_compliance int64 | compliance_reasoning string | compliance_elements_followed list | compliance_elements_missed list |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
In the tranquil solitude of a secluded chalet, Clyde meets an untimely death from snake venom; a case which sees the sharp mind of Detective Winston matching wits with two elusive suspects, Mack and Taylor.
Detective Winston was looking at the tailor-stitched files bundled together with papers on a desk in his office.... | Who is the most likely murderer? | ['Mack', 'Taylor'] | 0 | Mack | musr | [
{
"content": "You are a helpful assistant that will answer the questions given by the user.",
"role": "system"
},
{
"content": "IMPORTANT INSTRUCTION: Do NOT use a structured matrix, table, or grid approach to analyze the suspects. Do NOT create explicit rows/columns for means, motive, and opportuni... | [
"<think>\nFirst, I need to figure out who the murderer is between Mack and Taylor. The key is to check for means, motive, and opportunity. But I can't use a structured table; I have to reason naturally.\n\nLet me recall the details from the narrative.\n\nClyde was poisoned by snake venom in his chalet.\n\nNow, for ... | [
{
"endpoint_type": "chat",
"litellm_function": "acompletion",
"model": "together_ai/Qwen/Qwen3-Next-80B-A3B-Thinking",
"raw_prompt": [
{
"content": "You are a helpful assistant that will answer the questions given by the user.",
"role": "system"
},
{
"conten... | [
{
"cost": 0.0053664,
"finish_reason": "stop",
"input_tokens": 1426,
"model": "together_ai/Qwen/Qwen3-Next-80B-A3B-Thinking",
"output_tokens": 3435,
"raw_response": "{\"id\": \"oYRnQjF-3pDw3Z-9d20c745fd706a58\", \"created\": 1771787402, \"model\": \"together_ai/Qwen/Qwen3-Next-80B-A3B-Thinkin... | [
true
] | [
{
"correct_answer_index": 0,
"correct_answer_number": 1,
"correct_answer_text": "Mack",
"extracted_number": 1,
"extracted_text": "1",
"extractor": "musr",
"matched_by": "number",
"num_choices": 2,
"raw_line": "ANSWER: 1"
}
] | [
{
"endpoint_type": "chat",
"litellm_function": "acompletion",
"model": "gpt-4o-mini-2024-07-18",
"raw_prompt": [
{
"content": "You are an expert evaluator assessing whether a language model followed a prescribed reasoning strategy. You will be given:\n\n1. The STRATEGY INSTRUCTION that... | [
{
"cost": 0.0006657,
"finish_reason": "stop",
"input_tokens": 3922,
"model": "gpt-4o-mini-2024-07-18",
"output_tokens": 129,
"raw_response": "{\"id\": \"chatcmpl-DC98R86A5WijdkbvZXf1OGHkysnuY\", \"created\": 1771787423, \"model\": \"gpt-4o-mini-2024-07-18\", \"object\": \"chat.completion\", ... | 2 | The model acknowledges the need to evaluate means, motive, and opportunity but does not follow the prescribed criterion-first structure. It mixes evaluations of suspects and criteria, failing to create distinct sections for each analysis round as required by the strategy. | [
"SUSPECT INVENTORY",
"MEANS ANALYSIS",
"MOTIVE ANALYSIS",
"OPPORTUNITY ANALYSIS"
] | [
"CROSS-CUTTING COMPARISON",
"EXPLICIT LABELS",
"CRITERION-FIRST ORDER",
"SYNTHESIS"
] |
"In the dim light of an abandoned factory, Kaylee's lifeless body lay with the cold steel of a picka(...TRUNCATED) | Who is the most likely murderer? | ['Penelope', 'Francis'] | 0 | Penelope | musr | [{"content":"You are a helpful assistant that will answer the questions given by the user.","role":"(...TRUNCATED) | ["<think>\nFirst, I need to figure out who the murderer is between Penelope and Francis. The key is (...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/Qwen/Qwen3-Next-80B-A(...TRUNCATED) | [{"cost":0.00493695,"finish_reason":"stop","input_tokens":1483,"model":"together_ai/Qwen/Qwen3-Next-(...TRUNCATED) | [
true
] | [{"correct_answer_index":0,"correct_answer_number":1,"correct_answer_text":"Penelope","extracted_num(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"gpt-4o-mini-2024-07-18","raw_prom(...TRUNCATED) | [{"cost":0.00067515,"finish_reason":"stop","input_tokens":3933,"model":"gpt-4o-mini-2024-07-18","out(...TRUNCATED) | 1 | "The model completely ignores the prescribed strategy and does not follow the criterion-first struct(...TRUNCATED) | [] | ["SUSPECT INVENTORY","MEANS ANALYSIS","MOTIVE ANALYSIS","OPPORTUNITY ANALYSIS","CROSS-CUTTING COMPAR(...TRUNCATED) |
"In a shocking twist of events, Charlotte is found lifeless in her kitchen, felled by a stun gun, an(...TRUNCATED) | Who is the most likely murderer? | ['Letti', 'Deborah'] | 0 | Letti | musr | [{"content":"You are a helpful assistant that will answer the questions given by the user.","role":"(...TRUNCATED) | ["<think>\nFirst, I need to analyze the suspects: Letti and Deborah. The question is who is the most(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/Qwen/Qwen3-Next-80B-A(...TRUNCATED) | [{"cost":0.0043689,"finish_reason":"stop","input_tokens":1516,"model":"together_ai/Qwen/Qwen3-Next-8(...TRUNCATED) | [
true
] | [{"correct_answer_index":0,"correct_answer_number":1,"correct_answer_text":"Letti","extracted_number(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"gpt-4o-mini-2024-07-18","raw_prom(...TRUNCATED) | [{"cost":0.0006189,"finish_reason":"stop","input_tokens":3526,"model":"gpt-4o-mini-2024-07-18","outp(...TRUNCATED) | 1 | "The model completely disregards the prescribed reasoning strategy and does not follow the structure(...TRUNCATED) | [] | ["SUSPECT INVENTORY","MEANS ANALYSIS","MOTIVE ANALYSIS","OPPORTUNITY ANALYSIS","CROSS-CUTTING COMPAR(...TRUNCATED) |
"In the high-intensity world of a local fitness center, Tommy's exercise regime is abruptly terminat(...TRUNCATED) | Who is the most likely murderer? | ['Lucy', 'Jacqueline'] | 0 | Lucy | musr | [{"content":"You are a helpful assistant that will answer the questions given by the user.","role":"(...TRUNCATED) | ["<think>\nOkay, let's try to figure out who the murderer is here. So the victim is Tommy, killed wi(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/Qwen/Qwen3-Next-80B-A(...TRUNCATED) | [{"cost":0.0044401499999999995,"finish_reason":"stop","input_tokens":1451,"model":"together_ai/Qwen/(...TRUNCATED) | [
true
] | [{"correct_answer_index":0,"correct_answer_number":1,"correct_answer_text":"Lucy","extracted_number"(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"gpt-4o-mini-2024-07-18","raw_prom(...TRUNCATED) | [{"cost":0.0006345,"finish_reason":"stop","input_tokens":3578,"model":"gpt-4o-mini-2024-07-18","outp(...TRUNCATED) | 2 | "The model attempts to analyze the suspects but does not follow the prescribed criterion-first struc(...TRUNCATED) | [] | ["SUSPECT INVENTORY","MEANS ANALYSIS","MOTIVE ANALYSIS","OPPORTUNITY ANALYSIS","CROSS-CUTTING COMPAR(...TRUNCATED) |
"In a chilling turn of events, Rose is found lifeless in a car, killed by a vial of acid, leading De(...TRUNCATED) | Who is the most likely murderer? | ['Dexter', 'Daisy'] | 1 | Daisy | musr | [{"content":"You are a helpful assistant that will answer the questions given by the user.","role":"(...TRUNCATED) | ["<think>\nFirst, I need to figure out who the murderer is between Dexter and Daisy. The key is to c(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/Qwen/Qwen3-Next-80B-A(...TRUNCATED) | [{"cost":0.0047367,"finish_reason":"stop","input_tokens":1458,"model":"together_ai/Qwen/Qwen3-Next-8(...TRUNCATED) | [
true
] | [{"correct_answer_index":1,"correct_answer_number":2,"correct_answer_text":"Daisy","extracted_number(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"gpt-4o-mini-2024-07-18","raw_prom(...TRUNCATED) | [{"cost":0.0007000499999999999,"finish_reason":"stop","input_tokens":3727,"model":"gpt-4o-mini-2024-(...TRUNCATED) | 2 | "The model acknowledges the need to evaluate means, motive, and opportunity but does not follow the (...TRUNCATED) | ["Lists all suspects at the start (SUSPECT INVENTORY)","Includes YES/NO/UNCLEAR labels for each susp(...TRUNCATED) | ["Creates a dedicated MEANS ANALYSIS section evaluating ALL suspects on means (MEANS ROUND)","Create(...TRUNCATED) |
"In the dense, dank heart of the rainforest, Skylar's life was brutally silenced with a lead pipe, l(...TRUNCATED) | Who is the most likely murderer? | ['Albert', 'Mya'] | 1 | Mya | musr | [{"content":"You are a helpful assistant that will answer the questions given by the user.","role":"(...TRUNCATED) | ["<think>\nFirst, I need to figure out who the murderer is between Albert and Mya. The key points ar(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/Qwen/Qwen3-Next-80B-A(...TRUNCATED) | [{"cost":0.0045603,"finish_reason":"stop","input_tokens":1552,"model":"together_ai/Qwen/Qwen3-Next-8(...TRUNCATED) | [
true
] | [{"correct_answer_index":1,"correct_answer_number":2,"correct_answer_text":"Mya","extracted_number":(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"gpt-4o-mini-2024-07-18","raw_prom(...TRUNCATED) | [{"cost":0.0006217499999999999,"finish_reason":"stop","input_tokens":3625,"model":"gpt-4o-mini-2024-(...TRUNCATED) | 2 | "The model acknowledges the need to evaluate means, motive, and opportunity but does not follow the (...TRUNCATED) | [
"Motive analysis",
"Means analysis"
] | ["SUSPECT INVENTORY","MOTIVE ROUND","OPPORTUNITY ROUND","CROSS-CUTTING COMPARISON","EXPLICIT LABELS"(...TRUNCATED) |
"In the hallowed silence of a synagogue, a shocking murder unfolds as Derek is impaled by a harpoon,(...TRUNCATED) | Who is the most likely murderer? | ['Hailey', 'Angela'] | 1 | Angela | musr | [{"content":"You are a helpful assistant that will answer the questions given by the user.","role":"(...TRUNCATED) | ["<think>\nFirst, I need to figure out who the murderer is between Hailey and Angela. The key points(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/Qwen/Qwen3-Next-80B-A(...TRUNCATED) | [{"cost":0.0059913,"finish_reason":"stop","input_tokens":1812,"model":"together_ai/Qwen/Qwen3-Next-8(...TRUNCATED) | [
true
] | [{"correct_answer_index":1,"correct_answer_number":2,"correct_answer_text":"Angela","extracted_numbe(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"gpt-4o-mini-2024-07-18","raw_prom(...TRUNCATED) | [{"cost":0.00073845,"finish_reason":"stop","input_tokens":4335,"model":"gpt-4o-mini-2024-07-18","out(...TRUNCATED) | 1 | "The model completely ignores the prescribed reasoning strategy and does not follow the criterion-fi(...TRUNCATED) | [] | ["SUSPECT INVENTORY","MEANS ANALYSIS","MOTIVE ANALYSIS","OPPORTUNITY ANALYSIS","CROSS-CUTTING COMPAR(...TRUNCATED) |
"In the haze of neon lights and the serving of a silent hand of fate, Timothy lies dead in a casino,(...TRUNCATED) | Who is the most likely murderer? | ['Harry', 'Rosemary'] | 1 | Rosemary | musr | [{"content":"You are a helpful assistant that will answer the questions given by the user.","role":"(...TRUNCATED) | ["<think>\nFirst, I need to figure out who the murderer is between Harry and Rosemary. The key is to(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/Qwen/Qwen3-Next-80B-A(...TRUNCATED) | [{"cost":0.00477795,"finish_reason":"stop","input_tokens":1613,"model":"together_ai/Qwen/Qwen3-Next-(...TRUNCATED) | [
false
] | [{"correct_answer_index":1,"correct_answer_number":2,"correct_answer_text":"Rosemary","extracted_num(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"gpt-4o-mini-2024-07-18","raw_prom(...TRUNCATED) | [{"cost":0.0006472499999999999,"finish_reason":"stop","input_tokens":3759,"model":"gpt-4o-mini-2024-(...TRUNCATED) | 1 | "The model completely ignores the prescribed reasoning strategy and does not follow the structured a(...TRUNCATED) | [] | ["SUSPECT INVENTORY","MEANS ANALYSIS","MOTIVE ANALYSIS","OPPORTUNITY ANALYSIS","CROSS-CUTTING COMPAR(...TRUNCATED) |
"In the aftermath of a devastating home explosion that claims the life of Angela, seasoned Detective(...TRUNCATED) | Who is the most likely murderer? | ['Patrick', 'Alvin'] | 0 | Patrick | musr | [{"content":"You are a helpful assistant that will answer the questions given by the user.","role":"(...TRUNCATED) | ["<think>\nFirst, I need to figure out who the murderer is between Patrick and Alvin. I have to cons(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/Qwen/Qwen3-Next-80B-A(...TRUNCATED) | [{"cost":0.01224615,"finish_reason":"stop","input_tokens":1561,"model":"together_ai/Qwen/Qwen3-Next-(...TRUNCATED) | [
false
] | [{"correct_answer_index":0,"correct_answer_number":1,"correct_answer_text":"Patrick","extracted_numb(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"gpt-4o-mini-2024-07-18","raw_prom(...TRUNCATED) | [{"cost":0.0013768499999999998,"finish_reason":"stop","input_tokens":8603,"model":"gpt-4o-mini-2024-(...TRUNCATED) | 1 | "The model completely disregards the prescribed reasoning strategy and does not follow the structure(...TRUNCATED) | [] | ["SUSPECT INVENTORY","MEANS ANALYSIS","MOTIVE ANALYSIS","OPPORTUNITY ANALYSIS","CROSS-CUTTING COMPAR(...TRUNCATED) |
"In the frost-kissed tranquility of a chalet, Malaysia meets a fatal end by a rifle's sinister call,(...TRUNCATED) | Who is the most likely murderer? | ['Letti', 'Evelyn'] | 1 | Evelyn | musr | [{"content":"You are a helpful assistant that will answer the questions given by the user.","role":"(...TRUNCATED) | ["<think>\nFirst, I need to analyze the suspects: Letti and Evelyn. The question is who the most lik(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/Qwen/Qwen3-Next-80B-A(...TRUNCATED) | [{"cost":0.006438,"finish_reason":"stop","input_tokens":1380,"model":"together_ai/Qwen/Qwen3-Next-80(...TRUNCATED) | [
true
] | [{"correct_answer_index":1,"correct_answer_number":2,"correct_answer_text":"Evelyn","extracted_numbe(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"gpt-4o-mini-2024-07-18","raw_prom(...TRUNCATED) | [{"cost":0.00075975,"finish_reason":"stop","input_tokens":4549,"model":"gpt-4o-mini-2024-07-18","out(...TRUNCATED) | 2 | "The model acknowledges the strategy but does not follow the prescribed structure consistently. It e(...TRUNCATED) | [
"SUSPECT INVENTORY"
] | ["MEANS ANALYSIS","MOTIVE ANALYSIS","OPPORTUNITY ANALYSIS","CROSS-CUTTING COMPARISON","EXPLICIT LABE(...TRUNCATED) |
t1-strategy-musr-anti-qwen3-80b-thinking-musr-cf
Strategy compliance evaluation on MuSR murder mysteries — anti-strategy variant.
Model was explicitly told NOT to use matrix/table approach (anti-strategy control). Judge still scores against criterion-first rubric. Compliance is scored by an LLM judge (1-5 Likert) against the Criterion-First rubric.
Results
| Metric | Value |
|---|---|
| pass@1 | 0.8000 |
| Strategy compliance (mean) | 1.50 |
| Strategy compliance (min) | 1 |
| Strategy compliance (max) | 2 |
| Total problems | 10 |
Compliance Distribution
| Score | Count | Meaning |
|---|---|---|
| 1 | 5 | Non-compliant |
| 2 | 5 | Minimal |
| 3 | 0 | Partial |
| 4 | 0 | Mostly compliant |
| 5 | 0 | Fully compliant |
Details
| Parameter | Value |
|---|---|
| Model | together_ai/Qwen/Qwen3-Next-80B-A3B-Thinking |
| Thinking | True |
| Temperature | 0.7 |
| Max tokens | 32768 |
| Judge model | openai/gpt-4o-mini |
| Num examples | 10 |
| Variant | anti-strategy |
Usage
from datasets import load_dataset
ds = load_dataset("reasoning-degeneration-dev/t1-strategy-musr-anti-qwen3-80b-thinking-musr-cf", split="train")
scores = ds["strategy_compliance"]
print(f"Mean compliance: {sum(scores) / len(scores):.2f}")
Tracked in reasoning-degeneration-dev/PROJECT-MANIFEST
- Downloads last month
- 9