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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 128325
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on lambda‐style parameter tokens (the single‐letter identifiers immediately preceding the “=>” arrow in code).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    lambda expressions

    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
    Not in Any Lists

    No Comments

    Negative Logits
     some
    -1.59
     two
    -1.50
     at
    -1.48
     one
    -1.42
     several
    -1.40
     cesta
    -1.34
     almost
    -1.34
     used
    -1.33
     four
    -1.31
     three
    -1.30
    POSITIVE LOGITS
     '.')
    1.19
     horrid
    1.19
     quintessential
    1.16
     accéder
    1.15
    那样
    1.11
     molti
    1.10
     اکنون
    1.09
    è
    1.07
     établir
    1.05
     pacific
    1.05
    Activations Density 0.003%

    No Known Activations