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. 1168
    Prev
    Next
    INDEX
    Explanations

    phrases indicating the process of searching or discovering something

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

    terms related to persistence and consistency in performance

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

    find rhythm or momentum

    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
     '\\;'
    -0.74
    /**
    -0.71
     HasFactory
    -0.70
     يتيمه
    -0.64
     getRule
    -0.63
     kasarigan
    -0.63
    MigrationBuilder
    -0.63
    )++;
    -0.61
    IonicModule
    -0.59
     CreateTagHelper
    -0.59
    POSITIVE LOGITS
     momentum
    1.21
     rhythm
    1.10
    momentum
    0.91
    rhythm
    0.85
     groove
    0.83
     flow
    0.83
     ritmo
    0.79
     stride
    0.78
     Momentum
    0.75
     rhythms
    0.72
    Activations Density 0.052%

    No Known Activations