© 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. 13937
    Prev
    Next
    INDEX
    Explanations

    far as

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

    This neuron detects hedging phrases that start with “As far as I …,” such as “As far as I can tell” or “As far as I know.”

    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
     дуже
    -1.66
    órica
    -1.55
    庒
    -1.53
     土産
    -1.52
     semacam
    -1.52
     estadía
    -1.52
    噐
    -1.52
     墊
    -1.52
    襌
    -1.51
    猻
    -1.50
    POSITIVE LOGITS
    z
    2.02
    w
    2.02
    is
    1.95
    j
    1.93
    f
    1.80
     they
    1.77
    de
    1.73
    li
    1.73
    di
    1.70
     we
    1.68
    Activations Density 0.009%

    No Known Activations