© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 36292
    Prev
    Next
    INDEX
    Explanations

    sequences of true/false or numbers

    np_acts-logits-general · gemini-2.5-flash-lite

    This neuron fires on long, high-entropy alphanumeric tokens (e.g. base64-style strings, key/parameter blocks, or other random-looking numeric/letter sequences).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
     and
    -1.72
    ,
    -1.23
     kuh
    -1.17
     or
    -1.06
    DAFTAR
    -0.92
     from
    -0.92
     ОС
    -0.90
     They
    -0.89
     deceiving
    -0.88
    NRAS
    -0.88
    POSITIVE LOGITS
     prefeitura
    1.06
     Sólo
    1.05
     Tienes
    1.02
     eletrônico
    0.98
     Simplemente
    0.98
     Mientras
    0.96
     キ
    0.94
     Había
    0.93
     Muchos
    0.93
    aning
    0.93
    Activations Density 0.016%

    No Known Activations