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-2-2B
    4. Residual Stream - 65k
    5. 20-GEMMASCOPE-RES-65K
    6. 1001
    Prev
    Next
    INDEX
    Explanations

    references to social relationships

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

    references to relationships, particularly involving friends and family

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

    family, friends, colleagues

    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
     friend
    -1.59
    Friend
    -1.56
     Friend
    -1.55
    friend
    -1.55
     FRIEND
    -1.45
    FRIEND
    -1.37
     friendship
    -1.32
     Friends
    -1.22
     Friendship
    -1.22
     friends
    -1.16
    POSITIVE LOGITS
    ColumnHeaders
    0.57
    Portale
    0.55
     pleaſure
    0.55
     family
    0.53
     purpoſe
    0.53
     surla
    0.52
     collega
    0.52
     signal
    0.52
     colleagues
    0.50
     PhpStorm
    0.49
    Activations Density 0.014%

    No Known Activations