© 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. 22-GEMMASCOPE-RES-131K
    7. 127873
    Prev
    Next
    INDEX
    Explanations

    The main thing this neuron does is pick out section headers or title‐like labels (e.g. “[Indications]”, “Important information:”, “Size:”, “Product Description”) in structured text.

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408

    gizmos, pillowcases, TV stands, air, privacy, characters, preparing

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
     violento
    -1.21
     kalem
    -1.20
     praktis
    -1.16
     espl
    -1.15
     kontinu
    -1.15
     Eſ
    -1.13
    皷
    -1.13
     infrastruktur
    -1.13
    ERICAN
    -1.12
     mépris
    -1.12
    POSITIVE LOGITS
     with
    1.63
     like
    1.09
    ).
    1.05
     also
    1.03
    ".
    1.02
     =
    1.01
    ??
    1.00
     you
    0.98
     >
    0.98
     is
    0.97
    Activations Density 0.017%

    No Known Activations