© 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. 387
    Prev
    Next
    INDEX
    Explanations

    Friday afternoon, March, decline, engine, Chocolate

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

    The neuron fires on unusually specific or “rare” tokens—proper names, numbers, file or city identifiers, etc.—i.e. low-frequency tokens.

    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
     and
    -1.11
     véhic
    -1.11
     naviguant
    -1.04
     saisi
    -1.02
     commencent
    -0.98
     téléph
    -0.98
     rése
    -0.98
     comprennent
    -0.98
     Durée
    -0.96
     abſ
    -0.95
    POSITIVE LOGITS
     its
    1.02
     With
    0.95
     on
    0.94
     at
    0.89
     if
    0.88
    ált
    0.88
    ฒ
    0.86
     each
    0.82
     dorada
    0.82
     As
    0.82
    Activations Density 0.116%

    No Known Activations