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. 192
    Prev
    Next
    INDEX
    Explanations

    legal terminology related to criminal cases and trials

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

    terms related to criminal actions and legal proceedings

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

    committing crimes

    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
    Искәрмәләр
    -0.63
    الدراسه
    -0.63
    GIVEREF
    -0.61
    "}")
    -0.57
     дописавши
    -0.54
    DoubleQuotes
    -0.54
     препратки
    -0.54
     فريبيس
    -0.53
     substitutes
    -0.53
    Geplaatst
    -0.53
    POSITIVE LOGITS
     crimes
    1.10
     crime
    1.06
     murders
    0.85
    crimes
    0.79
     murder
    0.75
    crime
    0.75
     acts
    0.75
     atrocities
    0.72
     deeds
    0.67
     actions
    0.65
    Activations Density 0.186%

    No Known Activations