The dataset viewer is not available for this split.
Error code: FeaturesError Exception: ArrowInvalid Message: JSON parse error: Invalid value. in row 0 Traceback: Traceback (most recent call last): File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/packaged_modules/json/json.py", line 160, in _generate_tables df = pandas_read_json(f) File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/packaged_modules/json/json.py", line 38, in pandas_read_json return pd.read_json(path_or_buf, **kwargs) File "/src/services/worker/.venv/lib/python3.9/site-packages/pandas/io/json/_json.py", line 791, in read_json json_reader = JsonReader( File "/src/services/worker/.venv/lib/python3.9/site-packages/pandas/io/json/_json.py", line 905, in __init__ self.data = self._preprocess_data(data) File "/src/services/worker/.venv/lib/python3.9/site-packages/pandas/io/json/_json.py", line 917, in _preprocess_data data = data.read() File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/utils/file_utils.py", line 827, in read_with_retries out = read(*args, **kwargs) File "/usr/local/lib/python3.9/codecs.py", line 322, in decode (result, consumed) = self._buffer_decode(data, self.errors, final) UnicodeDecodeError: 'utf-8' codec can't decode byte 0xff in position 0: invalid start byte During handling of the above exception, another exception occurred: Traceback (most recent call last): File "/src/services/worker/src/worker/job_runners/split/first_rows.py", line 228, in compute_first_rows_from_streaming_response iterable_dataset = iterable_dataset._resolve_features() File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/iterable_dataset.py", line 3339, in _resolve_features features = _infer_features_from_batch(self.with_format(None)._head()) File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/iterable_dataset.py", line 2096, in _head return next(iter(self.iter(batch_size=n))) File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/iterable_dataset.py", line 2300, in iter for key, example in iterator: File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/iterable_dataset.py", line 1856, in __iter__ for key, pa_table in self._iter_arrow(): File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/iterable_dataset.py", line 1878, in _iter_arrow yield from self.ex_iterable._iter_arrow() File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/iterable_dataset.py", line 476, in _iter_arrow for key, pa_table in iterator: File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/iterable_dataset.py", line 323, in _iter_arrow for key, pa_table in self.generate_tables_fn(**gen_kwags): File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/packaged_modules/json/json.py", line 163, in _generate_tables raise e File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/packaged_modules/json/json.py", line 137, in _generate_tables pa_table = paj.read_json( File "pyarrow/_json.pyx", line 308, in pyarrow._json.read_json File "pyarrow/error.pxi", line 154, in pyarrow.lib.pyarrow_internal_check_status File "pyarrow/error.pxi", line 91, in pyarrow.lib.check_status pyarrow.lib.ArrowInvalid: JSON parse error: Invalid value. in row 0
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
PhysReason: A Comprehensive Benchmark towards Physics-Based Reasoning
PhysReason is accepted by ACL-2025-main
π Overview
PhysReason is a comprehensive physics-based reasoning benchmark consisting of 1,200 physics problems spanning multiple domains, with a focus on both knowledge-based (25%) and reasoning-based (75%) questions. This benchmark addresses the critical gap in evaluating large language models' capabilities in physics-based reasoning, which requires applying physics theorems and constraints in complex problem-solving scenarios.
β¨ Key Features
- π Dataset Size: 1,200 carefully curated physics problems
- π― Problem Types: Strategic mix of knowledge-based (25%) and reasoning-based (75%) questions
- π Theorem Coverage: Comprehensive coverage of 147 physics theorems
- π¨ Visual Content: 81% of problems include diagrams and visual elements
- π Difficulty Levels: Four distinct levels - Knowledge, Easy, Medium, Hard
- π Step-by-step Solutions: Average of 8.1 solution steps per problem (15.6 for hard problems)
- π Multi-modal: Supports both text and image inputs
π§ Data Collection
Our rigorous data collection process ensures high-quality, challenging problems:
- π Sources: Global college entrance exams and international physics competitions
- βοΈ Process: Standardized using MinerU framework for consistent formatting
- β Quality Control: Two-phase translation process with expert verification
- π Filtering: Systematically excluded easily searchable problems to prevent data leakage
- π Classification: Difficulty levels based on solving time and theorem complexity analysis
π Benchmark Comparison
Benchmark | Multi-modal | Size | Knowledge | Question Type | Avg. T | Step-by-step | Avg. T | Avg. S |
---|---|---|---|---|---|---|---|---|
JEEBench | β | 123 | CEE | OE,MC | 169.7 | - | - | - |
MMLU-Pro | β | 1299 | COL | MC | 52.1 | - | - | - |
GPQA | β | 227 | PH.D. | OE | 111.4 | β | 197.2 | 3.6 |
SciEval | β | 1657 | - | OE,MC | 154.5 | - | - | - |
SciBench | β | 295 | COL | OE | 80.5 | β | 315.9 | 2.8 |
MMMU | β | 443 | COL | OE,MC | 53.8 | - | - | - |
ScienceQA | β | 617 | K1-K12 | MC | 13.3 | β | 63.0 | 2.4 |
OlympiadBench | β | 2334 | COMP | OE | 222.0 | β | 199.8 | 3.7 |
EMMA | β | 156 | - | MC | 109.5 | - | - | - |
Ours-Knowledge | β | 300 | CEE+COMP | OE | 163.7 | β | 196.5 | 3.3 |
Ours-Easy | β | 300 | CEE+COMP | OE | 171.2 | β | 241.5 | 5.0 |
Ours-Medium | β | 300 | CEE+COMP | OE | 229.2 | β | 391.3 | 8.4 |
Ours-Hard | β | 300 | CEE+COMP | OE | 340.9 | β | 936.1 | 15.6 |
Ours-Full | β | 1200 | CEE+COMP | OE | 226.3 | β | 441.3 | 8.1 |
π Evaluation Framework
We introduce the Physics Solution Auto Scoring (PSAS) framework with two complementary evaluation approaches:
PSAS-A (Answer Level Evaluation)
- Sub-question Assessment: Evaluates answers for each sub-question independently
- LLM-based Extraction: Uses advanced language models for answer extraction
- Semantic Verification: Ensures semantic consistency between extracted and ground truth answers
- Weighted Scoring: Considers solution step lengths as weights for different sub-questions
PSAS-S (Step Level Evaluation)
Provides detailed step-by-step assessment through four phases:
- Data Extraction: Parses model responses and reference solutions
- Scoring: Evaluates correctness of each reasoning step
- First Error Detection: Identifies where models first deviate from correct reasoning
- Error Analysis: Classifies error types into four key bottlenecks:
- Physics Theorem Application
- Physics Process Understanding
- Calculation
- Physics Condition Analysis
π Usage
Core Evaluation Files
answer_evaluation_with_ds_ch_prompt.py
: Answer-level evaluation using Chinese promptsanswer_evaluation_with_ds_en_prompt.py
: Answer-level evaluation using English promptsformat_result_ds.py
: Optimizes unstable outputs into stable, consistent formatsstep_evaluation_with_ds_ch_prompt.py
: Step-level evaluation using Chinese promptsstep_evaluation_with_ds_en_prompt.py
: Step-level evaluation using English prompts
π Experimental Results
Non-O-like Models Performance
Model | Input | Knowledge | Easy | Medium | Hard | Avg. |
---|---|---|---|---|---|---|
Qwen2VL-72B | Q, I | 41.92/62.47 | 24.04/45.26 | 15.97/36.13 | 4.83/24.23 | 16.96/42.88 |
InternVL2.5-78B | Q, I | 28.34/64.71 | 24.16/50.69 | 17.72/38.56 | 9.71/25.95 | 19.98/45.89 |
GPT-4o | Q, I | 50.71/65.82 | 33.87/51.98 | 22.73/42.36 | 11.03/24.71 | 29.58/47.23 |
Deepseek-V3-671B | Q, IC | 55.86/66.14 | 40.06/52.77 | 26.63/44.02 | 13.73/26.87 | 34.07/48.42 |
Claude-3.5-Sonnet | Q, I | 54.14/66.45 | 41.35/55.85 | 28.14/44.86 | 15.11/28.51 | 34.69/49.88 |
Gemini-2.0-Flash | Q, I | 65.08/75.04 | 54.84/68.60 | 39.79/55.67 | 21.99/38.39 | 45.20/60.40 |
Gemini-2.0-Pro | Q, I | 67.99/79.01 | 55.43/71.47 | 44.29/57.74 | 23.81/42.66 | 47.88/62.74 |
O-like Models Performance
Model | Input | Knowledge | Easy | Medium | Hard | Avg. |
---|---|---|---|---|---|---|
o1-mini | Q, IC | 53.90/65.74 | 35.21/52.26 | 22.24/40.19 | 10.61/26.80 | 30.49/47.18 |
QvQ-72B | Q, I | 62.44/70.92 | 53.74/64.65 | 28.18/54.88 | 14.30/36.47 | 32.67/57.66 |
Gemini-2.0-Flash-Thinking-1206 | Q, I | 65.35/77.20 | 51.89/67.49 | 44.43/58.95 | 27.14/45.48 | 47.20/63.07 |
QwQ-32B | Q, IC | 62.03/76.28 | 54.92/71.08 | 43.64/62.14 | 22.99/42.19 | 45.89/63.87 |
GLM-Zero | Q, IC | 64.95/80.36 | 54.11/71.54 | 41.32/63.67 | 23.04/47.46 | 46.52/65.76 |
o3-mini-high | Q, IC | 70.67/83.61 | 67.20/81.95 | 45.31/64.57 | 30.12/47.23 | 53.32/69.34 |
Gemini-2.0-Flash-Thinking-0121 | Q, I | 73.44/84.15 | 63.17/75.94 | 50.41/66.60 | 31.90/48.47 | 54.73/69.73 |
Deepseek-R1 | Q, IC | 75.11/85.91 | 65.08/79.81 | 54.84/72.02 | 31.95/51.50 | 56.75/73.26 |
PhysReason-mini Results
Model | K. | E. | M. | H. | Avg. |
---|---|---|---|---|---|
o1-mini | 54.80 | 30.33 | 15.41 | 7.92 | 27.11 |
QvQ-72B | 51.17 | 37.10 | 29.83 | 22.13 | 35.06 |
QwQ-32B | 64.40 | 50.07 | 38.88 | 27.45 | 45.20 |
Gemini-2.0-Flash-Thinking-1206 | 71.47 | 49.97 | 36.83 | 22.97 | 45.42 |
GLM-Zero | 72.70 | 50.17 | 43.42 | 24.70 | 47.75 |
o1 | 72.47 | 53.37 | 49.31 | 25.32 | 50.12 |
o3-mini-high | 71.10 | 63.20 | 47.02 | 31.93 | 53.31 |
Gemini-2.0-Flash-Thinking-0121 | 76.33 | 56.87 | 51.85 | 32.61 | 54.42 |
Deepseek-R1 | 85.17 | 60.77 | 47.24 | 33.23 | 56.60 |
π Key Findings
- Performance Gap: Even top-performing models achieve less than 60% on answer-level evaluation
- Difficulty Scaling: Performance drops significantly from knowledge questions (75.11%) to hard problems (31.95%)
- O-like Model Advantage: Models with enhanced reasoning capabilities show superior performance
- Multi-modal Benefits: Visual content significantly enhances model understanding and performance
- Four Critical Bottlenecks identified through step-level evaluation:
- Physics Theorem Application
- Physics Process Understanding
- Calculation Accuracy
- Physics Condition Analysis
π Citation
If you find PhysReason useful in your research, please cite our paper:
@article{zhang2025physreason,
title={Physreason: A comprehensive benchmark towards physics-based reasoning},
author={Zhang, Xinyu and Dong, Yuxuan and Wu, Yanrui and Huang, Jiaxing and Jia, Chengyou and Fernando, Basura and Shou, Mike Zheng and Zhang, Lingling and Liu, Jun},
journal={arXiv preprint arXiv:2502.12054},
year={2025}
}
π License
This project is licensed under the MIT License - see the LICENSE file for details.
π§ Contact
We welcome contributions to PhysReason! Please contact us for more details.
π Quick Links:
- Downloads last month
- 203