Odd results in english when tabs/section sign § is present

[Fogapod]

use bm25::{EmbedderBuilder, Language, LanguageMode};

fn main() {
    let em = EmbedderBuilder::<u32>::with_avgdl(1.0)
        .b(0.0)
        .language_mode(LanguageMode::Fixed(Language::English))
        .build();

    dbg!(em.embed("§\t184.1240\tCarbon\tdioxide."));
    dbg!(em.embed("1\t184.1240\tCarbon\tdioxide."));
    dbg!(em.embed("\t184.1240\tCarbon\tdioxide."));
    dbg!(em.embed("§ 184.1240 Carbon dioxide."));
    dbg!(em.embed("184.1240 Carbon dioxide."));
}

results in

[src/main.rs:9:5] em.embed("§\t184.1240\tCarbon\tdioxide.") = Embedding(
    [
        TokenEmbedding {
            index: 504642266,
            value: 1.0,
        },
    ],
)
[src/main.rs:10:5] em.embed("1\t184.1240\tCarbon\tdioxide.") = Embedding(
    [
        TokenEmbedding {
            index: 2646367888,
            value: 1.0,
        },
        TokenEmbedding {
            index: 2840934837,
            value: 1.0,
        },
        TokenEmbedding {
            index: 196859388,
            value: 1.0,
        },
        TokenEmbedding {
            index: 2767204738,
            value: 1.0,
        },
    ],
)
[src/main.rs:11:5] em.embed("\t184.1240\tCarbon\tdioxide.") = Embedding(
    [
        TokenEmbedding {
            index: 2840934837,
            value: 1.0,
        },
        TokenEmbedding {
            index: 196859388,
            value: 1.0,
        },
        TokenEmbedding {
            index: 2767204738,
            value: 1.0,
        },
    ],
)
[src/main.rs:12:5] em.embed("§ 184.1240 Carbon dioxide.") = Embedding(
    [
        TokenEmbedding {
            index: 4201093589,
            value: 1.0,
        },
        TokenEmbedding {
            index: 2840934837,
            value: 1.0,
        },
        TokenEmbedding {
            index: 196859388,
            value: 1.0,
        },
        TokenEmbedding {
            index: 2767204738,
            value: 1.0,
        },
    ],
)
[src/main.rs:13:5] em.embed("184.1240 Carbon dioxide.") = Embedding(
    [
        TokenEmbedding {
            index: 2840934837,
            value: 1.0,
        },
        TokenEmbedding {
            index: 196859388,
            value: 1.0,
        },
        TokenEmbedding {
            index: 2767204738,
            value: 1.0,
        },
    ],
)

What's happening:

• having § and tabs results in a single token
• replacing § with other character results in correct tokens (even with tabs)
• replacing tabs with spaces results in correct tokens (even with §)

Is this expected?