© 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. 57673
    Prev
    Next
    INDEX
    Explanations

    phrases following commas

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

    This neuron is tracking each token’s absolute position in the document (i.e. it activates in proportion to the token’s running index).

    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
    carus
    -1.20
    gave
    -1.17
     takich
    -1.15
     WENT
    -1.09
    そんな
    -1.08
    黯
    -1.07
    ISER
    -1.05
    zzet
    -1.05
     และ
    -1.05
     اینکه
    -1.05
    POSITIVE LOGITS
     which
    5.50
    which
    2.53
     whose
    2.38
    ซึ่ง
    2.34
     designed
    2.13
     located
    2.02
     consisting
    2.02
     που
    1.93
     undertaken
    1.91
     described
    1.88
    Activations Density 0.190%

    No Known Activations