© 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. 0-GEMMASCOPE-TRANSCODER-16K
    7. 14520
    Prev
    Next
    INDEX
    Explanations

    verbs that describe actions affecting a noun, especially in the context of protection, reduction, or change

    oai_token-act-pair · gemini-2.0-flash

    prevention/minimization

    np_max-act · gemini-2.0-flash

    3rd-person singular verbs

    np_max-act-logits · gpt-5-miniTriggered by @chenshw0109
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_0/width_16k/average_l0_76
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.0.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
     teaches
    -1.67
     protects
    -1.63
     stimulates
    -1.60
     inspires
    -1.59
     motivates
    -1.58
     simplifies
    -1.58
     Efq
    -1.58
     eliminates
    -1.56
     prevents
    -1.56
     improves
    -1.55
    POSITIVE LOGITS
     the
    1.08
     a
    1.02
     on
    0.91
     in
    0.90
     an
    0.84
     to
    0.83
     some
    0.82
     one
    0.79
     more
    0.79
    ↵↵
    0.78
    Activations Density 3.858%

    No Known Activations