© 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. 35906
    Prev
    Next
    INDEX
    Explanations

    ways, including

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

    The neuron activates on nouns that denote categories or collections (e.g. “routes,” “reasons,” “variety,” “methods,” “applications,” “modalities”)—i.e. words used to enumerate types or groups.

    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
     or
    -2.39
     and
    -1.65
     this
    -1.46
     no
    -1.45
     would
    -1.44
     for
    -1.43
     only
    -1.43
     because
    -1.39
     will
    -1.38
     in
    -1.34
    POSITIVE LOGITS
    cluding
    1.59
    Including
    1.47
    INCLUDING
    1.39
     including
    1.37
     INCLUDING
    1.32
     havde
    1.24
     différents
    1.24
     différentes
    1.21
    including
    1.18
     chers
    1.17
    Activations Density 0.092%

    No Known Activations