Dalam kes anda, pilih elemen yang lebih khusus, iaitu pilih semua adik-beradik elemen elemen dengan nombor kelas:

soup.select('span.number ~ span')

Contoh

from bs4 import BeautifulSoup
html='''<p class="details">
<span>detail1</span>
<span class="number">1</span>
<span>detail2</span>
<span>detail3</span>
</p>'''
soup = BeautifulSoup(html)

[t.text for t in soup.select('span.number ~ span')]

Output

['detail2', 'detail3']

Anda boleh mencari semua <span> ​​dan melakukan pengindeksan biasa:

from bs4 import BeautifulSoup

html_doc = """\
<p class="details">
<span>detail1</span>
<span class="number">1</span>
<span>detail2</span>
<span>detail3</span>
</p>"""

soup = BeautifulSoup(html_doc, "html.parser")

spans = soup.find("p", class_="details").find_all("span")

for s in spans[-2:]:
    print(s.text)

Hasil keluaran:

detail2
detail3

Atau gunakan pemilih CSS:

spans = soup.select(".details span:nth-last-of-type(-n+2)")

for s in spans:
    print(s.text)

Hasil keluaran:

detail2
detail3