© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 22-TRANSCODER-HP
    7. 87952
    Prev
    Next
    INDEX
    Explanations

    have

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_22.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.22.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ersh
    -0.27
    —↵↵
    -0.27
    ç¨įå¾®
    -0.26
    ’ll
    -0.25
    ç¨į
    -0.25
    —I
    -0.25
    —he
    -0.24
     résult
    -0.24
    —in
    -0.23
    —is
    -0.23
    POSITIVE LOGITS
    æ¼Ĥ
    0.28
    æĦĪ
    0.25
     foot
    0.25
    ÑĢÑıд
    0.24
     reve
    0.24
    fts
    0.24
    ega
    0.24
     prow
    0.24
    çļĦç¡®æĺ¯
    0.23
    èģĭ
    0.23
    Activations Density 0.047%

    No Known Activations