© 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. 62404
    Prev
    Next
    INDEX
    Explanations

    hiding, concealing, or disguising

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

    The neuron is responding to high-frequency function words (common pronouns, auxiliaries, and similar “stop‐words” in English).

    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
     createdBy
    -0.74
    Advisory
    -0.72
    lln
    -0.69
    ัน
    -0.68
    اسی
    -0.66
    sistencia
    -0.65
    ئلة
    -0.65
    יסטוריה
    -0.65
     fekete
    -0.63
    رسی
    -0.63
    POSITIVE LOGITS
     hiding
    6.63
     hide
    6.25
     hides
    5.84
     hidden
    5.72
     hid
    5.44
    hide
    5.22
    hidden
    5.09
     Hiding
    5.06
    Hide
    5.06
    hiding
    4.97
    Activations Density 0.291%

    No Known Activations