据我所知，没有任何选项可用于设置一个值，当超过该值时会触发科学计数法。但是，您可以定义自己的字符串生成函数：

import numpy as np

def arr2str(arr, big=1e3, small_precision=0, big_precision=2):
    """Convert 1D array with floats to string.
    
    Parameters
    ----------
    arr : np.ndarray[float]
        To-be-converted array.
    big : float, optional
        Numbers for which the absolute value is >= `big` are considered big and
        will be printed in scientific notation. Otherwise the number is
        considered small and will not be printed in scientific notation.
    small_precision : float, optional
        Precision for small numbers.
    big_precision : float, optional
        Precision for big numbers.
    
    Returns
    -------
    str
    
    """
    # Check input validity
    if not arr.ndim == 1:
        raise ValueError("Array does not have a dimension of one")
    if not isinstance(arr[0], np.floating):
        raise ValueError("Array does not contain floats")
        
    str_list = []
    for ele in arr:
        if abs(ele) >= big:
            str_list.append(f'{ele:.{big_precision}e}')
        else:
            str_list.append(f'{ele:.{small_precision}f}')
    string = ' '.join(str_list)
    string = '[' + string + ']'
    return string

# Usage example
example_arr = np.array([-80., -75., -70., -65., 55., -50., 4294, -64921])
print(arr2str(example_arr))

打印：

[-80 -75 -70 -65 55 -50 4.29e+03 -6.49e+04]