© 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
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 57411
    Prev
    Next
    INDEX
    Explanations

    I got

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

    The neuron strongly activates on first-person self-references—pronouns like “I,” “me,” “my,” and “myself”—marking personal narrator statements.

    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
    他们
    -1.09
     jugadores
    -1.05
     their
    -0.97
     JTextField
    -0.96
    ׂ
    -0.96
    ͏
    -0.91
    niem
    -0.91
    玩家
    -0.89
    oncesto
    -0.89
    icted
    -0.88
    POSITIVE LOGITS
     my
    1.35
     feels
    1.20
     feeling
    1.15
     felt
    1.10
     mistakes
    1.01
    策略
    1.00
     conditions
    0.99
     feel
    0.95
     conserve
    0.95
     tactics
    0.94
    Activations Density 0.012%

    No Known Activations