© 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. 39218
    Prev
    Next
    INDEX
    Explanations

    idiomatic phrases and descriptions

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

    The neuron strongly activates on words that express spatial or directional relations (e.g. under, back, down, up)

    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
     of
    -1.39
    苅
    -1.23
    !!!!!!!
    -1.20
     !!!!
    -1.20
    öpfer
    -1.19
    nhung
    -1.19
    !!!!!!
    -1.17
    jaket
    -1.14
     continually
    -1.13
    成为了
    -1.13
    POSITIVE LOGITS
     plantas
    1.34
     with
    1.32
     There
    1.28
     membersihkan
    1.28
     merawat
    1.26
     any
    1.25
     soothing
    1.23
    🥽
    1.23
     if
    1.19
     áreas
    1.17
    Activations Density 0.049%

    No Known Activations