© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 5244
    Prev
    Next
    INDEX
    Explanations

    'might be', 'aware that', 'following closely', 'finally found', 'talked about'

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

    The neuron fires on first-person self-references (tokens like “I”, “I’m”, “I’ve”, etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ﻠ
    0.67
    Ꭳ
    0.57
    colm
    0.56
    ﻤ
    0.56
    canic
    0.55
    ﻨ
    0.55
    ல்கள்
    0.55
    IELD
    0.55
    <0x83>
    0.54
    景
    0.54
    POSITIVE LOGITS
     myself
    0.74
     pribadi
    0.70
     jestem
    0.66
     faccio
    0.61
     unwittingly
    0.59
     pegawai
    0.59
     dieses
    0.58
     defeats
    0.58
     persönlich
    0.58
     hastily
    0.58
    Activations Density 0.085%

    No Known Activations