Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 92735
    Prev
    Next
    INDEX
    Explanations

    concepts related to barriers, limitations, and transitions between different states or realms

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

    boundaries, barriers, and breaches

    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-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
    parsedMessage
    -0.54
    BoxShadow
    -0.52
    ########.
    -0.51
    bootstrapcdn
    -0.48
    SourceChecksum
    -0.46
    rentiel
    -0.45
    statistic
    -0.43
     Palt
    -0.43
    rencies
    -0.42
    enschaften
    -0.42
    POSITIVE LOGITS
     ujednoznacz
    0.74
    webElementXpaths
    0.46
     barrier
    0.39
    Tikang
    0.39
     barriers
    0.36
     åp
    0.36
    Hozzáférés
    0.35
     border
    0.35
    writerow
    0.34
    HttpPut
    0.34
    Activations Density 0.100%

    No Known Activations