© 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. 22-GEMMASCOPE-TRANSCODER-16K
    7. 16367
    Prev
    Next
    INDEX
    Explanations

    clauses with "was," "were," "had," or other forms of "be" describing locations, actions, and relationships of people and things.

    oai_token-act-pair · gemini-2.0-flash

    was were

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_22/width_16k/average_l0_15
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.22.ln2.hook_normalized
    Architecture
    jumprelu_transcoder
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    SourceChecksum
    -0.60
    äuser
    -0.57
     referenties
    -0.56
    getDoctrine
    -0.56
    LocationManager
    -0.53
    üge
    -0.52
     SqlDbType
    -0.51
    hões
    -0.51
    المشاركات
    -0.50
     Administrativna
    -0.50
    POSITIVE LOGITS
     being
    0.65
    standing
    0.61
     standing
    0.60
    
    0.57
     AssemblyCompany
    0.56
     packing
    0.55
     combing
    0.55
     balancing
    0.54
     washing
    0.54
     playing
    0.54
    Activations Density 1.841%

    No Known Activations