© 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. 34-GEMMASCOPE-RES-131K
    7. 3018
    Prev
    Next
    INDEX
    Explanations

    Japanese media and programming

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

    The neuron fires on tokens containing uppercase letters—e.g. proper names, titles, HTML tags, camelCase identifiers, or any token with a capital character.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     Szab
    -0.84
     Sert
    -0.82
     
    -0.80
     Twilight
    -0.76
     twilight
    -0.75
    meris
    -0.74
     sò
    -0.73
     sabe
    -0.72
    experience
    -0.72
    смотрим
    -0.72
    POSITIVE LOGITS
     Saitama
    1.24
    Vil
    0.84
     superman
    0.80
     Imo
    0.80
    琦
    0.79
    Ces
    0.76
     gravely
    0.76
     welder
    0.76
     Igbo
    0.75
    fdata
    0.73
    Activations Density 0.005%

    No Known Activations