© 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. 22-GEMMASCOPE-RES-131K
    4. 28121
    Prev
    Next
    INDEX
    Explanations

    type, associated, actually

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

    The neuron is primarily detecting numeric tokens (i.e. standalone numbers and numerical values).

    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
     constater
    -1.52
     ſol
    -1.46
     corriger
    -1.38
     zoude
    -1.34
     remé
    -1.34
     diriger
    -1.27
     télécharger
    -1.26
     offrir
    -1.24
     compléter
    -1.21
    
    -1.21
    POSITIVE LOGITS
     while
    1.55
     more
    1.43
     as
    1.41
     their
    1.40
     when
    1.39
     there
    1.36
     When
    1.31
    When
    1.29
     went
    1.28
     all
    1.27
    Activations Density 0.014%

    No Known Activations