© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 64934
    Prev
    Next
    INDEX
    Explanations

    adjectives and their qualities

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

    The neuron detects subjective evaluative adjectives (like “beautiful,” “scary,” “funny,” “cool,” etc.) that express judgment or opinion.

    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
    T
    -1.31
    ˃
    -1.25
    ことができます
    -1.21
    S
    -1.18
    L
    -1.18
    TE
    -1.13
    NC
    -1.09
    ↵↵
    -1.09
    Aa
    -1.09
     ggf
    -1.09
    POSITIVE LOGITS
     or
    1.39
     seguirá
    1.38
     when
    1.36
    もありますが
    1.32
    tipation
    1.25
     wohnen
    1.23
     according
    1.22
     dará
    1.21
    elé
    1.20
     convencer
    1.20
    Activations Density 0.062%

    No Known Activations