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-2-2B
    4. Residual Stream - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 14605
    Prev
    Next
    INDEX
    Explanations

    elements related to individual choices and beliefs regarding religion and marriage

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

    Follows a word like "one", "is", "will"

    np_acts-logits-general · gemini-2.0-flash

    deciding or court judgments

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Partial
    -0.44
     partial
    -0.44
     obligé
    -0.43
    olid
    -0.43
    Partial
    -0.42
    Xna
    -0.41
    bete
    -0.40
    backing
    -0.40
     importance
    -0.39
    ことがあります
    -0.39
    POSITIVE LOGITS
     decide
    1.16
     decides
    1.13
     decisions
    1.03
     Decide
    1.02
     deciding
    1.01
     decision
    0.98
    decide
    0.97
    Decide
    0.95
     décider
    0.92
     Decisions
    0.88
    Activations Density 0.352%

    No Known Activations