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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 73625
    Prev
    Next
    INDEX
    Explanations

    information related to identity and connections associated with individuals or entities

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

    Words or phrases that establish connections or relationships between concepts, entities, or ideas. This includes terms that indicate identification, connection, attribution, discovery, or relation between elements (like "with", "between", "associated with", "attributed", "connect the dots", "author", "identification", "was", "discovery", "related to").

    eleuther_acts_top20 · claude-3-7-sonnet-20250219Triggered by @lu-christina
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114
    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
    CompleteListener
    -0.51
    HasBeenSet
    -0.47
    yntaxException
    -0.42
    ListItemIcon
    -0.40
    consulté
    -0.40
    ScreenState
    -0.38
    afficheront
    -0.37
    wapV
    -0.36
    kém
    -0.36
     Statement
    -0.36
    POSITIVE LOGITS
     contextLoads
    0.52
    berdayakan
    0.51
     سكانية
    0.50
    ViewFeatures
    0.49
     identific
    0.47
     stanovnika
    0.46
     jScrollPane
    0.43
    fromnode
    0.43
    Identyfik
    0.42
    juana
    0.42
    Activations Density 0.143%

    No Known Activations