© 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-2-2B
    5. 18-RES-MATRYOSHKA-DC
    7. 127
    Prev
    Next
    INDEX
    Explanations

    the word "which" and nearby helper verbs such as "are", "can", "will", and "have"

    oai_token-act-pair · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    chanind/gemma-2-2b-batch-topk-matryoshka-saes-w-32k-l0-40/standard/blocks.18.hook_resid_post
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    chanind/pile-uncopyrighted-gemma-1024-abbrv-1B
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     scolaires
    -0.60
    Chosen
    -0.60
     snowing
    -0.56
     skrevet
    -0.53
     legyen
    -0.50
     storming
    -0.48
     asfalto
    -0.48
     kidding
    -0.48
     électroniques
    -0.47
    age
    -0.46
    POSITIVE LOGITS
     would
    0.97
     could
    0.96
     is
    0.96
     might
    0.96
     can
    0.96
     may
    0.92
     فريبيس
    0.89
     will
    0.88
    ագրություններ
    0.84
     unfortunately
    0.82
    Activations Density 1.092%

    No Known Activations