© 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. 73225
    Prev
    Next
    INDEX
    Explanations

    programming file paths

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

    The neuron fires on code‐related path and filename tokens (e.g. directory names, file basenames, extensions) as they appear in source‐file and build script lines.

    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
     on
    -1.61
     while
    -1.57
     indicated
    -1.46
     earlier
    -1.39
     великолеп
    -1.38
     during
    -1.38
     last
    -1.38
     this
    -1.37
     young
    -1.34
     indicate
    -1.34
    POSITIVE LOGITS
    ſa
    1.55
    蝼
    1.55
     thoſe
    1.54
     triomphe
    1.53
    Especially
    1.45
    توضیحات
    1.45
     preguntó
    1.44
     Eſ
    1.43
    erobic
    1.42
    Ferien
    1.42
    Activations Density 0.060%

    No Known Activations