© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 118969
    Prev
    Next
    INDEX
    Explanations

    back door

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

    This neuron fires on directional particles like “out” (especially in phrasal verbs such as “build out” or “sneak out”) and “back” (as in “back door”), i.e. words that signal movement or covert routing.

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     nữa
    -0.83
    mydb
    -0.80
    jComboBox
    -0.79
    支店
    -0.72
    цами
    -0.70
    說是
    -0.69
     utanför
    -0.69
    étranger
    -0.68
    zner
    -0.68
    防水
    -0.67
    POSITIVE LOGITS
     back
    4.84
    Back
    3.22
    back
    3.16
     Back
    3.00
     BACK
    2.75
    BACK
    2.64
    バック
    2.27
     rear
    1.91
     назад
    1.84
     зад
    1.80
    Activations Density 0.025%

    No Known Activations