© 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. 12797
    Prev
    Next
    INDEX
    Explanations

    elements related to account management, including usernames and credentials

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

    mentions and values of authentication/secret credentials in code or configuration (e.g., API/auth data embedded in env vars, headers, or connection strings).

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    authentication credentials and API keys in code, particularly in variable assignments and configuration contexts.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    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
    errHandler
    -0.52
    documentElement
    -0.48
    iastical
    -0.48
    uteria
    -0.47
    ILON
    -0.46
    ogia
    -0.46
    SuspendLayout
    -0.46
    delwed
    -0.45
    ValueStyle
    -0.45
    cshtml
    -0.45
    POSITIVE LOGITS
     cred
    0.40
     token
    0.37
     credentials
    0.36
     linkovi
    0.36
    TryDecodeAsNil
    0.35
     passwords
    0.35
     kasarigan
    0.34
     password
    0.34
    OGND
    0.34
     tambin
    0.34
    Activations Density 0.096%

    No Known Activations