© 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
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 3967
    Prev
    Next
    INDEX
    Explanations

    special markers and delimiters in the text, particularly at the beginning of sections or with high prioritization

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

    special tokens or markers at the beginning of a sequence (like <bos>).

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    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
    AISSEE
    -0.51
    parsedMessage
    -0.50
    mployment
    -0.47
    principalColumn
    -0.47
    Биография
    -0.46
     Wikiseite
    -0.45
     समीक्षाओं
    -0.45
     Głów
    -0.44
    DeleteBehavior
    -0.44
    EndContext
    -0.43
    POSITIVE LOGITS
     dante
    0.49
    Cow
    0.47
    нская
    0.44
     Cow
    0.44
     Stony
    0.43
    :+:
    0.43
    εβ
    0.43
    ɔ
    0.42
    Suns
    0.42
    KELEY
    0.42
    Activations Density 0.260%

    No Known Activations