Why needlebench summarizer hardcodes 4k in the result demonstration?

[b2r66sun]

I found the needlebench summarizer hardcodes 4K in the multi-needle-reasoning, no matter what dataset_size is given, e.g. the 2-Needle-EN-4K in the last several lines of code. Is it a typo or the 4K does not mean the dataset size?

summarizer_config = {
        'type': NeedleBenchSummarizer,
        'summary_groups': summary_groups,
        'dataset_abbrs': [
            'NeedleBench-Overall-Score',
            f'--------- NeedleBench-{dataset_size.upper()}-Single-Needle-Retrieval ---------',
            'Single-Needle-Retrieval(S-RT)',
            'Single-Needle-Retrieval-EN',
            'Single-Needle-Retrieval-ZH',
            f'--------- NeedleBench-{dataset_size.upper()}-Multi-Needle-Retrieval ---------',
            'Multi-Needle-Retrieval(M-RT)',
            'Multi-Needle-Retrieval-EN',
            'Multi-Needle-Retrieval-ZH',
            f'--------- NeedleBench-{dataset_size.upper()}-Multi-Needle-Reasoning ---------',
            'Multi-Needle-Reasoning(M-RS)',
            'Multi-Needle-Reasoning-EN',
            'Multi-Needle-Reasoning-ZH',
            '2-Needle-EN-4K',
            '2-Needle-ZH-4K',
            '3-Needle-EN-4K',
            '3-Needle-ZH-4K',
            '4-Needle-EN-4K',
            '4-Needle-ZH-4K',
            '5-Needle-EN-4K',
            '5-Needle-ZH-4K',
            ]
        }

[DseidLi]

Thank you for bringing this to my attention. Upon review, it's indeed a typo. I'll fix it promptly. Much appreciated for your attention to detail!

[DseidLi]

Solved in #1125

[b2r66sun]

I see. Thanks