© 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. 3315
    Prev
    Next
    INDEX
    Explanations

    the words "latter" and "former", and sometimes other words depending on context

    oai_token-act-pair · gemini-2.0-flash

    reference to previously mentioned

    np_max-act · gemini-2.0-flash

    latter Method used: 1 — because the top tokens are predominantly "latter"

    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
     Offisielt
    -0.81
    <bos>
    -0.73
    دانشنامهٔ
    -0.71
     faſt
    -0.71
     Efq
    -0.67
     raiſ
    -0.64
    ſelf
    -0.64
     whoſe
    -0.64
     juſt
    -0.64
    ormais
    -0.61
    POSITIVE LOGITS
    uanya
    0.58
     CreateTagHelper
    0.58
    addContainerGap
    0.57
    AttributeSet
    0.56
    
    0.55
    adaptiveStyles
    0.54
     Catawiki
    0.54
     متعلقه
    0.53
    共
    0.53
    Slack
    0.51
    Activations Density 1.333%

    No Known Activations