© 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. 33416
    Prev
    Next
    INDEX
    Explanations

    links to flickr content like secret, id, server attributes

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

    The neuron specializes in spotting Flickr API (or XML/json) photo‐metadata tokens—things like “photo,” “id,” “secret,” “server,” “farm,” “exif,” “flickr,” and related numeric IDs.

    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
     Bildschirm
    -1.31
     norvég
    -1.24
     bolsos
    -1.24
     comprimidos
    -1.21
    玏
    -1.18
     regulation
    -1.17
    魑
    -1.17
     angeb
    -1.16
    ISION
    -1.16
     или
    -1.16
    POSITIVE LOGITS
     it
    1.30
     you
    1.18
    Altre
    1.16
     Конечно
    1.13
    比較
    1.11
     Потому
    1.05
    argento
    1.04
    それほど
    1.01
     twill
    1.00
     bevorzug
    0.99
    Activations Density 0.004%

    No Known Activations