© 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. 10-GEMMASCOPE-RES-131K
    7. 73863
    Prev
    Next
    INDEX
    Explanations

    announcing, inviting, suggesting, thanking

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

    The neuron fires on formal statement verbs in press‐release–style language (e.g. “announces,” “apologizes,” “invites,” “commits,” “thanks”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     about
    -1.55
     recently
    -1.40
     exactly
    -1.32
     OGSÅ
    -1.28
     of
    -1.25
     exemplaires
    -1.23
     recurso
    -1.21
    ᅚ
    -1.20
     especialistas
    -1.20
     sólo
    -1.20
    POSITIVE LOGITS
    because
    1.41
     their
    1.39
    釦
    1.36
    𝑼
    1.34
     Because
    1.23
    是因为
    1.20
     hObject
    1.20
    是因為
    1.18
     люди
    1.16
    !'
    1.14
    Activations Density 0.123%

    No Known Activations