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": "HERE IS YOUR REQUIRED REASONING STRATEGY: You must solve this murder mystery using COUNTERFACTUAL HYPOTHESIS TESTING WITH EARLY TERMINATION. This strategy works by assuming... | [
"STEP 1 – SUSPECT INVENTORY \n1. Taylor \n2. Mack \n\nSTEP 2 – HYPOTHESIS TESTING \n\nHYPOTHESIS: Taylor is the killer. \nCONDITION 1 (Means): Taylor must have had access to snake venom. \nEvidence: She owns a cleaning business that services sites with venomous snakes, has taken a reptile-handling/first-aid c... | [
{
"endpoint_type": "chat",
"litellm_function": "acompletion",
"model": "together_ai/moonshotai/Kimi-K2-Instruct-0905",
"raw_prompt": [
{
"content": "You are a helpful assistant that will answer the questions given by the user.",
"role": "system"
},
{
"conten... | [
{
"cost": 0.003758,
"finish_reason": "stop",
"input_tokens": 2099,
"model": "together_ai/moonshotai/Kimi-K2-Instruct-0905",
"output_tokens": 553,
"raw_response": "{\"id\": \"oYRjWB8-zqrih-9d20b96e2dd73d08\", \"created\": 1771786826, \"model\": \"together_ai/moonshotai/Kimi-K2-Instruct-0905\"... | [
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": 0.00036885,
"finish_reason": "stop",
"input_tokens": 1843,
"model": "gpt-4o-mini-2024-07-18",
"output_tokens": 154,
"raw_response": "{\"id\": \"chatcmpl-DC8yoCfX3ZW3hxQYpsPDNKQrmbMUg\", \"created\": 1771786826, \"model\": \"gpt-4o-mini-2024-07-18\", \"object\": \"chat.completion\",... | 5 | The model fully complies with the prescribed strategy by systematically listing suspects, framing hypotheses, deriving necessary conditions, and testing them sequentially with explicit verdicts. It adheres to the early termination rule and provides clear hypotheses verdicts for each suspect, demonstrating a thorough un... | [
"SUSPECT INVENTORY",
"HYPOTHESIS FRAMING",
"CONDITIONS FIRST",
"SEQUENTIAL TESTING",
"EXPLICIT VERDICTS",
"EARLY TERMINATION",
"HYPOTHESIS VERDICT",
"EFFICIENCY"
] | [] |
In the dim light of an abandoned factory, Kaylee's lifeless body lay with the cold steel of a pickaxe lodged within, thrusting Detective Winston into a dark game of deception with suspects Penelope and Francis.
Detective Winston sat at the precinct and scrutinized the last seen diary of Kaylee, a murder victim. The di... | 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": "system"
},
{
"content": "HERE IS YOUR REQUIRED REASONING STRATEGY: You must solve this murder mystery using COUNTERFACTUAL HYPOTHESIS TESTING WITH EARLY TERMINATION. This strategy works by assuming... | [
"STEP 1 – SUSPECT INVENTORY \n1. Penelope \n2. Francis \n\nSTEP 2 – HYPOTHESIS TESTING \n\nHYPOTHESIS: Penelope is the killer. \n\nCONDITION 1 (Means): Penelope must have had access to the pickaxe used as the murder weapon. \nEvidence: A neighbor saw Penelope “with a pickaxe a few times recently around the sa... | [
{
"endpoint_type": "chat",
"litellm_function": "acompletion",
"model": "together_ai/moonshotai/Kimi-K2-Instruct-0905",
"raw_prompt": [
{
"content": "You are a helpful assistant that will answer the questions given by the user.",
"role": "system"
},
{
"conten... | [
{
"cost": 0.003202,
"finish_reason": "stop",
"input_tokens": 2152,
"model": "together_ai/moonshotai/Kimi-K2-Instruct-0905",
"output_tokens": 350,
"raw_response": "{\"id\": \"oYRjWAa-4qdXgU-9d20b96e2cc1f5f6\", \"created\": 1771786816, \"model\": \"together_ai/moonshotai/Kimi-K2-Instruct-0905\... | [
true
] | [
{
"correct_answer_index": 0,
"correct_answer_number": 1,
"correct_answer_text": "Penelope",
"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": 0.0003429,
"finish_reason": "stop",
"input_tokens": 1658,
"model": "gpt-4o-mini-2024-07-18",
"output_tokens": 157,
"raw_response": "{\"id\": \"chatcmpl-DC8yoPNAMG94EcBzRp2DfoctPT1xh\", \"created\": 1771786826, \"model\": \"gpt-4o-mini-2024-07-18\", \"object\": \"chat.completion\", ... | 5 | The model fully complies with the prescribed strategy by listing all suspects, framing each as a hypothesis, and deriving necessary conditions before testing them sequentially. It uses explicit verdicts for each condition and applies early termination correctly, providing a clear verdict for each suspect. The structure... | [
"SUSPECT INVENTORY",
"HYPOTHESIS FRAMING",
"CONDITIONS FIRST",
"SEQUENTIAL TESTING",
"EXPLICIT VERDICTS",
"EARLY TERMINATION",
"HYPOTHESIS VERDICT",
"EFFICIENCY"
] | [] |
In a shocking twist of events, Charlotte is found lifeless in her kitchen, felled by a stun gun, and it's up to seasoned Detective Winston to unravel the mystery, with prime suspects being Letti and Deborah.
Winston lit up a fresh cigarette, the inky night a silent observer as he reviewed Charlotte's file. It wasn't ... | 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": "system"
},
{
"content": "HERE IS YOUR REQUIRED REASONING STRATEGY: You must solve this murder mystery using COUNTERFACTUAL HYPOTHESIS TESTING WITH EARLY TERMINATION. This strategy works by assuming... | [
"STEP 1 — SUSPECT INVENTORY \n1. Deborah \n2. Letti \n\nSTEP 2 — HYPOTHESIS TESTING \n\nHYPOTHESIS: Deborah is the killer. \nCONDITION 1 (Means): Deborah must have had access to a stun-gun. \nEvidence: Narrative never places a stun-gun in Deborah’s hands or shows her buying one. \nVerdict: CONTRADICTED \n\n... | [
{
"endpoint_type": "chat",
"litellm_function": "acompletion",
"model": "together_ai/moonshotai/Kimi-K2-Instruct-0905",
"raw_prompt": [
{
"content": "You are a helpful assistant that will answer the questions given by the user.",
"role": "system"
},
{
"conten... | [
{
"cost": 0.003162,
"finish_reason": "stop",
"input_tokens": 2181,
"model": "together_ai/moonshotai/Kimi-K2-Instruct-0905",
"output_tokens": 327,
"raw_response": "{\"id\": \"oYRjWB2-zqrih-9d20b96e38b60aae\", \"created\": 1771786819, \"model\": \"together_ai/moonshotai/Kimi-K2-Instruct-0905\"... | [
true
] | [
{
"correct_answer_index": 0,
"correct_answer_number": 1,
"correct_answer_text": "Letti",
"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": 0.0003369,
"finish_reason": "stop",
"input_tokens": 1630,
"model": "gpt-4o-mini-2024-07-18",
"output_tokens": 154,
"raw_response": "{\"id\": \"chatcmpl-DC8yocPBFh0w6uJa21ta6TMjXWiJg\", \"created\": 1771786826, \"model\": \"gpt-4o-mini-2024-07-18\", \"object\": \"chat.completion\", ... | 5 | The model fully complies with the prescribed strategy by listing all suspects, framing each as a hypothesis, and deriving necessary conditions before testing them. It also uses explicit verdicts, applies early termination correctly, and provides a clear verdict for each suspect. The structure and flow of reasoning alig... | [
"SUSPECT INVENTORY",
"HYPOTHESIS FRAMING",
"CONDITIONS FIRST",
"SEQUENTIAL TESTING",
"EXPLICIT VERDICTS",
"EARLY TERMINATION",
"HYPOTHESIS VERDICT",
"EFFICIENCY"
] | [] |
"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) | ["STEP 1 – SUSPECT INVENTORY \n1. Lucy \n2. Jacqueline \n\nSTEP 2 – HYPOTHESIS TESTING \n\nH(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/moonshotai/Kimi-K2-In(...TRUNCATED) | [{"cost":0.0031279999999999997,"finish_reason":"stop","input_tokens":2117,"model":"together_ai/moons(...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":0.00033659999999999994,"finish_reason":"stop","input_tokens":1640,"model":"gpt-4o-mini-2024(...TRUNCATED) | 5 | "The model fully complies with the prescribed strategy by listing all suspects, framing each as a hy(...TRUNCATED) | ["SUSPECT INVENTORY","HYPOTHESIS FRAMING","CONDITIONS FIRST","SEQUENTIAL TESTING","EXPLICIT VERDICTS(...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) | ["STEP 1 – SUSPECT INVENTORY \n1. Dexter \n2. Daisy \n\nSTEP 2 – HYPOTHESIS TESTING \n\nHYPO(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/moonshotai/Kimi-K2-In(...TRUNCATED) | [{"cost":0.0037489999999999997,"finish_reason":"stop","input_tokens":2126,"model":"together_ai/moons(...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":0.00036195,"finish_reason":"stop","input_tokens":1841,"model":"gpt-4o-mini-2024-07-18","out(...TRUNCATED) | 5 | "The model fully complies with the prescribed strategy by systematically listing suspects, framing h(...TRUNCATED) | ["SUSPECT INVENTORY","HYPOTHESIS FRAMING","CONDITIONS FIRST","SEQUENTIAL TESTING","EXPLICIT VERDICTS(...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) | ["STEP 1 — SUSPECT INVENTORY \n1. Mya \n2. Albert \n\nSTEP 2 — HYPOTHESIS TESTING \n\nHYPOTH(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/moonshotai/Kimi-K2-In(...TRUNCATED) | [{"cost":0.003203,"finish_reason":"stop","input_tokens":2225,"model":"together_ai/moonshotai/Kimi-K2(...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":0.0003396,"finish_reason":"stop","input_tokens":1636,"model":"gpt-4o-mini-2024-07-18","outp(...TRUNCATED) | 5 | "The model fully complies with the prescribed strategy by listing all suspects, framing each as a hy(...TRUNCATED) | ["SUSPECT INVENTORY","HYPOTHESIS FRAMING","CONDITIONS FIRST","SEQUENTIAL TESTING","EXPLICIT VERDICTS(...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) | ["STEP 1 – SUSPECT INVENTORY \n1. Hailey \n2. Angela \n\nSTEP 2 – HYPOTHESIS TESTING \n\nHYP(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/moonshotai/Kimi-K2-In(...TRUNCATED) | [{"cost":0.003864,"finish_reason":"stop","input_tokens":2478,"model":"together_ai/moonshotai/Kimi-K2(...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":0.00035925,"finish_reason":"stop","input_tokens":1771,"model":"gpt-4o-mini-2024-07-18","out(...TRUNCATED) | 5 | "The model fully complies with the prescribed strategy by clearly listing all suspects, framing each(...TRUNCATED) | ["SUSPECT INVENTORY","HYPOTHESIS FRAMING","CONDITIONS FIRST","SEQUENTIAL TESTING","EXPLICIT VERDICTS(...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) | ["STEP 1 – SUSPECT INVENTORY \n1. Harry \n2. Rosemary \n\nSTEP 2 – HYPOTHESIS TESTING \n\nHY(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/moonshotai/Kimi-K2-In(...TRUNCATED) | [{"cost":0.003548,"finish_reason":"stop","input_tokens":2285,"model":"together_ai/moonshotai/Kimi-K2(...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":0.00035415,"finish_reason":"stop","input_tokens":1721,"model":"gpt-4o-mini-2024-07-18","out(...TRUNCATED) | 5 | "The model fully complies with the prescribed strategy by listing all suspects, framing each as a hy(...TRUNCATED) | ["SUSPECT INVENTORY","HYPOTHESIS FRAMING","CONDITIONS FIRST","SEQUENTIAL TESTING","EXPLICIT VERDICTS(...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) | ["STEP 1 — SUSPECT INVENTORY \n1. Patrick \n2. Alvin \n\nSTEP 2 — HYPOTHESIS TESTING \n\nHYP(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/moonshotai/Kimi-K2-In(...TRUNCATED) | [{"cost":0.003445,"finish_reason":"stop","input_tokens":2221,"model":"together_ai/moonshotai/Kimi-K2(...TRUNCATED) | [
true
] | [{"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":0.00035055,"finish_reason":"stop","input_tokens":1709,"model":"gpt-4o-mini-2024-07-18","out(...TRUNCATED) | 5 | "The model fully complies with the prescribed strategy by listing all suspects, framing each as a hy(...TRUNCATED) | ["SUSPECT INVENTORY","HYPOTHESIS FRAMING","CONDITIONS FIRST","SEQUENTIAL TESTING","EXPLICIT VERDICTS(...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) | ["STEP 1 – SUSPECT INVENTORY \n1. Evelyn \n2. Letti \n\nSTEP 2 – HYPOTHESIS TESTING \n\nHYPO(...TRUNCATED) | [{"endpoint_type":"chat","litellm_function":"acompletion","model":"together_ai/moonshotai/Kimi-K2-In(...TRUNCATED) | [{"cost":0.00317,"finish_reason":"stop","input_tokens":2042,"model":"together_ai/moonshotai/Kimi-K2-(...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":0.00033765,"finish_reason":"stop","input_tokens":1675,"model":"gpt-4o-mini-2024-07-18","out(...TRUNCATED) | 5 | "The model fully complies with the prescribed strategy by listing all suspects, framing each as a hy(...TRUNCATED) | ["SUSPECT INVENTORY","HYPOTHESIS FRAMING","CONDITIONS FIRST","SEQUENTIAL TESTING","EXPLICIT VERDICTS(...TRUNCATED) | [] |
t1-strategy-musr-counterfactual-kimi-k2-instruct-musr-cfact
Strategy compliance evaluation on MuSR murder mysteries — counterfactual-hypothesis variant.
Model was instructed to use Counterfactual Hypothesis Testing with Early Termination (assume each suspect is guilty, test conditions, stop at first contradiction). Compliance is scored by an LLM judge (1-5 Likert) against the Criterion-First rubric.
Results
| Metric | Value |
|---|---|
| pass@1 | 0.9000 |
| Strategy compliance (mean) | 5.00 |
| Strategy compliance (min) | 5 |
| Strategy compliance (max) | 5 |
| Total problems | 10 |
Compliance Distribution
| Score | Count | Meaning |
|---|---|---|
| 1 | 0 | Non-compliant |
| 2 | 0 | Minimal |
| 3 | 0 | Partial |
| 4 | 0 | Mostly compliant |
| 5 | 10 | Fully compliant |
Details
| Parameter | Value |
|---|---|
| Model | together_ai/moonshotai/Kimi-K2-Instruct-0905 |
| Thinking | False |
| Temperature | 0.7 |
| Max tokens | 32768 |
| Judge model | openai/gpt-4o-mini |
| Num examples | 10 |
| Variant | counterfactual-hypothesis |
Usage
from datasets import load_dataset
ds = load_dataset("reasoning-degeneration-dev/t1-strategy-musr-counterfactual-kimi-k2-instruct-musr-cfact", 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
- 8