© 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-3-27B-IT
    5. 22-GEMMASCOPE-2-TRANSCODER-262K
    7. 42218
    Prev
    Next
    INDEX
    Explanations

    constants in code

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/transcoder_all/layer_22_width_262k_l0_small_affine
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     udrž
    0.41
     находятся
    0.41
     eleven
    0.40
    in
    0.39
    rom
    0.39
     균
    0.39
     giãn
    0.39
        
    0.38
     duet
    0.38
    avu
    0.38
    POSITIVE LOGITS
    아가
    0.45
    গুলোকে
    0.42
     Clipboard
    0.41
     tactics
    0.41
     Switching
    0.41
    茆
    0.41
    ويل
    0.40
    AgentError
    0.40
     Pooling
    0.40
     copying
    0.38
    Activations Density 0.001%

    No Known Activations