© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 130146
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on content-bearing nouns (especially technical or domain-specific terms).

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

    instead or rather

    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
     or
    -0.99
     crouching
    -0.84
    したのが
    -0.83
    即是
    -0.82
     raíz
    -0.81
    alloon
    -0.81
    술
    -0.80
     could
    -0.80
     indiscriminate
    -0.80
     by
    -0.79
    POSITIVE LOGITS
     instead
    2.31
    instead
    2.11
    Instead
    1.81
     zamiast
    1.73
     вместо
    1.68
     Instead
    1.62
     invece
    1.58
     rather
    1.27
    prefer
    1.22
    место
    1.22
    Activations Density 0.180%

    No Known Activations