MuSR

Overview

MuSR (Multistep Soft Reasoning) is a benchmark for evaluating complex reasoning abilities through narrative-based problems. It includes murder mysteries, object placements, and team allocation scenarios requiring multi-step inference.

Task Description

Task Type: Complex Reasoning (Multiple-Choice)
Input: Narrative scenario with question and answer choices
Output: Correct answer letter (A-F)
Domains: Murder mysteries, object tracking, team allocation

Key Features

Narrative-based reasoning problems
Requires multi-step logical inference
Three distinct reasoning domains
Tests constraint satisfaction and deduction
Longer context requiring careful reasoning

Evaluation Notes

Default configuration uses 0-shot evaluation
Uses Chain-of-Thought (CoT) prompting
Three subsets: murder_mysteries, object_placements, team_allocation
Simple accuracy metric
Challenging benchmark requiring careful reading

Properties

Property	Value
Benchmark Name	`musr`
Dataset ID	AI-ModelScope/MuSR
Paper	N/A
Tags	`MCQ`, `Reasoning`
Metrics	`acc`
Default Shots	0-shot
Evaluation Split	`test`

Data Statistics

Metric	Value
Total Samples	756
Prompt Length (Mean)	4891.57 chars
Prompt Length (Min/Max)	2812 / 7537 chars

Per-Subset Statistics:

Subset	Samples	Prompt Mean	Prompt Min	Prompt Max
`murder_mysteries`	250	5743.1	4056	7537
`object_placements`	256	5294.0	3735	7525
`team_allocation`	250	3627.93	2812	4351

Sample Example

Subset: murder_mysteries

{
  "input": [
    {
      "id": "5ec1a7bd",
      "content": "Answer the following multiple choice question. The last line of your response should be of the following format: 'ANSWER: [LETTER]' (without quotes) where [LETTER] is one of A,B. Think step by step before answering.\n\nIn an adrenaline inducing ... [TRUNCATED] ... and wronged, over and over, at the same sight. It was quite a sight. \n\nWinston, shuffling back to the station, was left with one thought - Looks like Mackenzie had quite an eventful week.\n\nWho is the most likely murderer?\n\nA) Mackenzie\nB) Ana"
    }
  ],
  "choices": [
    "Mackenzie",
    "Ana"
  ],
  "target": "A",
  "id": 0,
  "group_id": 0
}

Note: Some content was truncated for display.

Prompt Template

Prompt Template:

Answer the following multiple choice question. The last line of your response should be of the following format: 'ANSWER: [LETTER]' (without quotes) where [LETTER] is one of {letters}. Think step by step before answering.

{question}

{choices}

Usage

Using CLI

evalscope eval \
    --model YOUR_MODEL \
    --api-url OPENAI_API_COMPAT_URL \
    --api-key EMPTY_TOKEN \
    --datasets musr \
    --limit 10  # Remove this line for formal evaluation

Using Python

from evalscope import run_task
from evalscope.config import TaskConfig

task_cfg = TaskConfig(
    model='YOUR_MODEL',
    api_url='OPENAI_API_COMPAT_URL',
    api_key='EMPTY_TOKEN',
    datasets=['musr'],
    dataset_args={
        'musr': {
            # subset_list: ['murder_mysteries', 'object_placements', 'team_allocation']  # optional, evaluate specific subsets
        }
    },
    limit=10,  # Remove this line for formal evaluation
)

run_task(task_cfg=task_cfg)

3.5 KiB Raw Blame History