© 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. Gemma-2-2B
    5. 0-GEMMASCOPE-TRANSCODER-16K
    7. 3971
    Prev
    Next
    INDEX
    Explanations

    the word "be", particularly when followed by function words such as "to."

    oai_token-act-pair · gemini-2.0-flash

    be

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_0/width_16k/average_l0_76
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.0.ln2.hook_normalized
    Architecture
    jumprelu_transcoder
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     be
    -1.04
     Be
    -0.83
    be
    -0.69
     J
    -0.64
    <eos>
    -0.63
     $
    -0.61
    Be
    -0.61
     BE
    -0.60
     K
    -0.60
     C
    -0.59
    POSITIVE LOGITS
     Efq
    1.55
     itſelf
    1.41
     myſelf
    1.39
     Jefus
    1.34
     houſe
    1.33
     Majefty
    1.27
     pleaſure
    1.27
     raiſ
    1.27
     themſelves
    1.26
     Houſe
    1.24
    Activations Density 0.525%

    No Known Activations