© 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. 33-TRANSCODER-HP
    7. 54079
    Prev
    Next
    INDEX
    Explanations

    punctuation and suffixes

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_33.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.33.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    )test
    -0.30
    AtA
    -0.29
    ertest
    -0.29
     ÑĢеÑģÑĤ
    -0.27
     Provid
    -0.27
     Jog
    -0.26
    @student
    -0.26
    "is
    -0.25
    following
    -0.25
    toa
    -0.25
    POSITIVE LOGITS
    -s
    0.59
    ,
    0.45
    -S
    0.43
    :
    0.40
    .
    0.40
    ï¼Į
    0.38
    _S
    0.38
    ä¹ĭç±»çļĦ
    0.36
    .S
    0.36
    _s
    0.35
    Activations Density 0.041%

    No Known Activations