© 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. 34-GEMMASCOPE-RES-131K
    7. 3224
    Prev
    Next
    INDEX
    Explanations

    Avengers movies and characters

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

    The neuron strongly activates on multi‐word, title‐style proper names—especially blockbuster movie titles and franchise names (e.g. “Captain America: Civil War,” “Avengers Infinity War,” “MCU,” etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    myra
    -0.80
     ביט
    -0.73
    彬
    -0.73
    untamiento
    -0.67
     Partei
    -0.66
    umbar
    -0.65
    监
    -0.65
     though
    -0.64
    arkan
    -0.63
    サイ
    -0.63
    POSITIVE LOGITS
     Endgame
    1.00
     Captain
    0.98
     Marvel
    0.91
     Avengers
    0.91
     Russo
    0.89
    Avengers
    0.88
    Captain
    0.84
    Shield
    0.83
     Shield
    0.81
    Politik
    0.80
    Activations Density 0.022%

    No Known Activations