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-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 5636
    Prev
    Next
    INDEX
    Explanations

    references to animal limbs or appendages, particularly claws and paws

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

    paw, claw, spur

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    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
    TextAppearance
    -0.47
    basicConfig
    -0.46
     transport
    -0.44
     فريبيس
    -0.44
    TemporalType
    -0.43
    ymce
    -0.43
    ricts
    -0.42
     الرياضيه
    -0.41
     Cet
    -0.40
     Transport
    -0.39
    POSITIVE LOGITS
     Paw
    1.00
    Paw
    0.99
     paw
    0.88
    paw
    0.87
     Claw
    0.84
     Paws
    0.80
     claw
    0.75
    Claw
    0.73
     Pawn
    0.70
     paws
    0.69
    Activations Density 0.008%

    No Known Activations