© 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. 65630
    Prev
    Next
    INDEX
    Explanations

    action verbs followed by prepositions

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

    The neuron fires on bits of personal context or “back‐story” statements in a post—things like “I had a discussion,” “I’ve been thinking,” “I tried,” “I came across,” etc., i.e. expressions of the author’s own actions or experiences.

    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
    dienne
    -1.27
    ⛽
    -1.18
    还没有
    -1.18
     Eğer
    -1.15
    vska
    -1.12
    plication
    -1.09
    nment
    -1.09
    yya
    -1.09
    ifers
    -1.09
    tiers
    -1.09
    POSITIVE LOGITS
     lots
    1.23
     couldn
    1.20
     gotta
    1.19
     biggest
    1.11
     mantiene
    1.08
     thru
    1.07
     didn
    1.05
    땐
    1.05
     ativar
    1.03
     дают
    1.02
    Activations Density 0.118%

    No Known Activations