© 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. 73013
    Prev
    Next
    INDEX
    Explanations

    designed feature descriptions

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

    The neuron lights up on very common English function words (e.g. for, to, the, and) rather than content words.

    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
     poliuret
    -1.50
    ländische
    -1.39
    ureusement
    -1.36
     edifici
    -1.31
    şı
    -1.31
    sieme
    -1.30
     chrétien
    -1.28
    Į
    -1.27
     patrie
    -1.27
    Unnamed
    -1.24
    POSITIVE LOGITS
     our
    1.88
     your
    1.25
    のですか
    1.23
     tantangan
    1.23
     ponsel
    1.22
     jembatan
    1.20
     homicidio
    1.20
     ofrecemos
    1.18
     bawat
    1.17
     durch
    1.13
    Activations Density 0.247%

    No Known Activations