© 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. 10578
    Prev
    Next
    INDEX
    Explanations

    The neuron seems to be detecting the word "form" or "own", as well as other words with relatively low consistency

    oai_token-act-pair · gemini-2.0-flash

    own

    np_max-act · gemini-2.0-flash

    own

    np_max-act-logits · gpt-5-miniTriggered by @chenshw0109
    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
    oltre
    -0.74
    none
    -0.73
    <bos>
    -0.71
     raiſ
    -0.69
    own
    -0.68
     ролях
    -0.67
    now
    -0.65
    all
    -0.65
    DockStyle
    -0.63
    less
    -0.62
    POSITIVE LOGITS
     less
    0.77
     more
    0.65
     nej
    0.62
     weaker
    0.60
    reya
    0.60
    GroupLayout
    0.59
     giuri
    0.58
     naj
    0.57
    rind
    0.57
     les
    0.57
    Activations Density 3.597%

    No Known Activations