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-2-2B
    4. Residual Stream - 65k
    5. 20-GEMMASCOPE-RES-65K
    6. 252
    Prev
    Next
    INDEX
    Explanations

    instances and references to the word "shoot" and its variations related to violent actions

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

    terms related to shooting incidents

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

    shoot, shooting, shot

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    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
    theless
    -0.83
    ]<<"
    -0.80
     Дата
    -0.76
    "}")
    -0.70
    GIVEREF
    -0.68
    帖最后由
    -0.68
     Deuter
    -0.66
    "}"
    -0.66
    ----</
    -0.65
    TagMode
    -0.64
    POSITIVE LOGITS
    ंदीखरीदारी
    0.67
    outs
    0.66
     shoots
    0.65
    öll
    0.65
    nuff
    0.64
    gun
    0.63
    pad
    0.61
    guns
    0.60
    poietic
    0.59
    تفصیلات
    0.58
    Activations Density 0.015%

    No Known Activations