© 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. 58002
    Prev
    Next
    INDEX
    Explanations

    words before negative concepts

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

    The neuron isn’t picking out any particular word or meaning but rather simply tracks how far into the text you are: its activation grows with a token’s position in the sequence.

    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
     basic
    -1.13
    pantalón
    -0.95
    Proble
    -0.95
     nuclear
    -0.91
    景點
    -0.87
     behaviour
    -0.85
    basic
    -0.84
     normal
    -0.84
     FontAwesome
    -0.82
    PROBLE
    -0.81
    POSITIVE LOGITS
    ﴿
    0.93
    低
    0.92
     toujours
    0.90
    吁
    0.86
    (!_
    0.85
    terer
    0.84
     preocupaciones
    0.83
    బ
    0.83
     امروز
    0.83
     distillery
    0.83
    Activations Density 0.021%

    No Known Activations