© 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. 21-GEMMASCOPE-TRANSCODER-16K
    7. 7393
    Prev
    Next
    INDEX
    Explanations

    instances of the fragment "ma" that are often followed by programming- or service-related terms like "gento", "preduce", "ppers", "elstrom", "ight", or "service"

    oai_token-act-pair · gemini-2.0-flash

    tokens starting with 'ma'

    np_token-act-pair-logits · gpt-4o-mini

    Code/configuration snippets

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_21/width_16k/average_l0_13
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.21.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
    mybatisplus
    -0.73
     GenerationType
    -0.61
     ModelExpression
    -0.60
     Résultats
    -0.57
    TagMode
    -0.56
     Doi
    -0.56
     iNdEx
    -0.56
    UVWXYZ
    -0.54
     powied
    -0.51
     Socialista
    -0.50
    POSITIVE LOGITS
    ]=>
    0.51
    DATES
    0.45
    #+#
    0.44
     pen
    0.43
    schemas
    0.43
    はじめに
    0.43
     воро
    0.43
    oredCriteria
    0.43
    KEYCODE
    0.42
     Matta
    0.42
    Activations Density 1.456%

    No Known Activations