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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 16k
    5. 9-GEMMASCOPE-RES-16K
    6. 3237
    Prev
    Next
    INDEX
    Explanations

    instances of the word "protect" and related concepts of safeguarding or defense

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

    safeguarding or defense

    np_acts-logits-general · gemini-2.0-flash

    protect

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_9/width_16k/average_l0_88
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.9.hook_resid_post
    Hook Layer
    9
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    SequentialGroup
    -0.54
    inSlope
    -0.54
     createdBy
    -0.50
    ponential
    -0.48
     dequeue
    -0.47
    +#+
    -0.47
     AspNetCore
    -0.47
     versa
    -0.46
    raisemb
    -0.46
     PopupWindow
    -0.45
    POSITIVE LOGITS
     protect
    2.06
     protection
    2.00
    protect
    1.92
     Protect
    1.88
     protecting
    1.87
    Protect
    1.84
     Protection
    1.84
    Protection
    1.82
    protection
    1.81
     protects
    1.80
    Activations Density 0.116%

    No Known Activations