OpenCompass now supports the brand new long-context language model evaluation benchmark — [BABILong](https://arxiv.org/pdf/2406.10149). BABILong provides an evaluation of long-context reasoning across extremely long documents, including a diverse set of 20 reasoning tasks such as fact chaining, simple induction, deduction, counting, and handling lists/sets. This benchmark is designed to test the ability of language models to reason over facts distributed in long natural text, and it allows for the construction of tasks of almost arbitrary length to adapt to the evaluation of new, more powerful models in an extensible and controllable way.
## How to Use
The BABILong dataset is available on Hugging Face: [RMT-team/babilong](https://huggingface.co/datasets/RMT-team/babilong). Opencompass provides an automatic download for BABILong dataset, due to the dataset size, we only provide the data up to 1M tokens. For longer context, you can download the dataset from Hugging Face directly.
BABILong paper provides in total 20 tasks, we provide 10 tasks configurations in OpenCompass and they are organized by different context sizes. You can create your own configurations by following the examples in `opencompass/configs/datasets/babilong/babilong_1m_gen.py`.
Opencompass provides a demo for evaluating language models on the BABILong dataset.
```bash
opencompass configs/eval_babilong.py
```
OpenCompass provides the results of some models on the BABILong dataset. The evaluation results are run with LMDeploy with default model settings.
Please as a grading expert, judge whether the final answers given by the candidates below are consistent with the standard answers, that is, whether the candidates answered correctly.
Here are some evaluation criteria:
1. Please refer to the given standard answer. You don't need to re-generate the answer to the question because the standard answer has been given. You only need to judge whether the candidate's answer is consistent with the standard answer according to the form of the question. Don't try to answer the original question. You can assume that the standard answer is definitely correct.
2. Because the candidate's answer may be different from the standard answer in the form of expression, before making a judgment, please understand the question and the standard answer first, and then judge whether the candidate's answer is correct, but be careful not to try to answer the original question.
3. Some answers may contain multiple items, such as multiple-choice questions, multiple-select questions, fill-in-the-blank questions, etc. As long as the answer is the same as the standard answer, it is enough. For multiple-select questions and multiple-blank fill-in-the-blank questions, the candidate needs to answer all the corresponding options or blanks correctly to be considered correct.
4. Some answers may be expressed in different ways, such as some answers may be a mathematical expression, some answers may be a textual description, as long as the meaning expressed is the same. And some formulas are expressed in different ways, but they are equivalent and correct.
5. If the prediction is given with \\boxed{}, please ignore the \\boxed{} and only judge whether the candidate's answer is consistent with the standard answer.
Please judge whether the following answers are consistent with the standard answer based on the above criteria. Grade the predicted answer of this new question as one of:
A: CORRECT
B: INCORRECT
Just return the letters "A" or "B", with no text around it.
Here is your task. Simply reply with either CORRECT, INCORRECT. Don't apologize or correct yourself if there was a mistake; we are just trying to grade the answer.
f"Follow the given examples and answer the question.\n{_hint}\n\nQ: {{input}}\nA: Let's think step by step. And you must give your final answer by starting with 'So the answer is' "
f"Follow the given examples and answer the question.\n{_hint}\n\nQ: {{input}}\nA: Let's think step by step. And you must give your final answer by starting with 'So the answer is' "
