© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 128692
    Prev
    Next
    INDEX
    Explanations

    The neuron is tuned to detect promotional and digital‐media marketing terms (e.g. free, download, best, official, buy, album, PDF).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    item followed by description

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     only
    -1.80
     then
    -1.66
     their
    -1.36
     like
    -1.36
    豬
    -1.23
     every
    -1.22
     barely
    -1.16
     more
    -1.16
     your
    -1.13
     they
    -1.13
    POSITIVE LOGITS
    Different
    1.43
     aprobó
    1.33
    importante
    1.30
    Mutability
    1.30
    different
    1.28
     также
    1.27
     olika
    1.24
    異なる
    1.23
    Aufbau
    1.23
     désolés
    1.21
    Activations Density 0.002%

    No Known Activations