© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 5051
    Prev
    Next
    INDEX
    Explanations

    This neuron detects phrases that refer to “interests” (e.g. best interests, legitimate interest) or similar self-serving/legal‐justification wording.

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
    陪伴
    0.55
    一带
    0.54
    汥
    0.54
    をや
    0.52
     አበባ
    0.51
    豔
    0.51
    neyland
    0.50
    齐全
    0.50
     acompaña
    0.50
    汗
    0.50
    POSITIVE LOGITS
     interests
    2.23
    interests
    2.16
     Interests
    2.06
     kepentingan
    1.98
     selfish
    1.95
    利益
    1.93
     intereses
    1.90
     Interessen
    1.79
     interesses
    1.79
     স্বার্থ
    1.78
    Activations Density 0.546%

    No Known Activations