© 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. 10785
    Prev
    Next
    INDEX
    Explanations

    abstract words related to goals or intention

    oai_token-act-pair · gemini-2.0-flash

    aim

    np_max-act · gemini-2.0-flash

    aim Method used: 1 — because every max activating token is "▁aim" — aim

    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
     the
    -1.04
    ↵↵
    -0.68
     a
    -0.65
     both
    -0.59
     all
    -0.57
     his
    -0.56
    %@",
    -0.55
     I
    -0.55
     efforts
    -0.54
     it
    -0.53
    POSITIVE LOGITS
     виправивши
    0.90
    +#+#
    0.84
     nakalista
    0.76
     متعلقه
    0.75
     termica
    0.74
     Numa
    0.74
    WithEmail
    0.73
     Benfica
    0.71
     CSIRO
    0.70
     Etr
    0.70
    Activations Density 6.215%

    No Known Activations