© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 9457
    Prev
    Next
    INDEX
    Explanations

    choosing the bestreplacement

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

    The neuron strongly activates on code identifiers for types or classes—e.g. named types like Event, Runnable, Model, Class, Task, etc.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
    Not in Any Lists

    No Comments

    Negative Logits
     Whitehall
    -0.90
     Jehovah
    -0.87
    DOU
    -0.85
    ABBRE
    -0.84
    kelen
    -0.81
     उन
    -0.81
    ̈r
    -0.81
    她们
    -0.80
    czegó
    -0.80
     wilde
    -0.79
    POSITIVE LOGITS
     chosen
    0.93
     behind
    0.91
     replacement
    0.91
    <0x96>
    0.90
     it
    0.89
     that
    0.85
    Replacement
    0.81
    replacement
    0.81
    Replace
    0.81
     best
    0.80
    Activations Density 0.085%

    No Known Activations