© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 91008
    Prev
    Next
    INDEX
    Explanations

    buying drugs and sexual content

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

    The neuron fires on prominent, domain‐specific topic headings or titles (e.g. “Diabetes,” “Sexual abuse of children,” “Buy Cheap,” “Research Paper,” “Tips on How to Make Your Penis Stay Hard Longer!”)—in other words, it spotlights section or document titles and headlines.

    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
     on
    -1.32
     their
    -1.24
     one
    -1.22
     at
    -1.20
     after
    -1.16
     terribly
    -1.15
     utterly
    -1.15
     difficult
    -1.14
     these
    -1.13
     those
    -1.13
    POSITIVE LOGITS
     AMAZING
    1.42
     REALLY
    1.41
    trying
    1.23
     Herbs
    1.23
     exuber
    1.22
     EVEN
    1.22
    VERY
    1.16
    
    1.16
    wiście
    1.16
     extravag
    1.14
    Activations Density 0.001%

    No Known Activations