© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 4800
    Prev
    Next
    INDEX
    Explanations

    phrases related to suggestions and recommendations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    text of a conversation or presentation where recommendations or suggestions are being given.

    oai_token-act-pair · gemini-2.0-flashTriggered by @ywhp2c8tjm

    sentences that deliver strong advice, recommendations, or warnings.

    oai_token-act-pair · gpt-4.1-2025-04-14Triggered by @ywhp2c8tjm

    Following recommendations or advice

    np_acts-logits-general · gemini-2.0-flash

    recommendation and advice

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    asteroide
    -0.43
    libft
    -0.38
    metheus
    -0.38
     Ehre
    -0.36
    onedDateTime
    -0.36
    曖昧さ回避
    -0.36
    willigung
    -0.35
     argint
    -0.35
     ffilmiau
    -0.35
     textStatus
    -0.34
    POSITIVE LOGITS
     recommendations
    0.77
     Recommendations
    0.73
     recommendation
    0.73
    advice
    0.72
    recommendations
    0.71
     recomenda
    0.70
     recommended
    0.69
     raccomand
    0.69
     recommends
    0.69
    recommendation
    0.69
    Activations Density 0.613%

    No Known Activations