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. 1016
    Prev
    Next
    INDEX
    Explanations

    sentence-ending punctuation marks and function calls

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

    punctuation marks and specific function call patterns

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

    string/list manipulation

    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
    Jegyzetek
    -0.52
    RegressionTest
    -0.50
    alised
    -0.48
    modelBuilder
    -0.44
    UserScript
    -0.44
     Staf
    -0.44
    SYNOPSIS
    -0.44
     thiệu
    -0.43
    AutoScale
    -0.43
    Javadoc
    -0.43
    POSITIVE LOGITS
     nakalista
    0.66
     EconPapers
    0.61
     '\\;'
    0.57
    agrama
    0.56
    tvguidetime
    0.56
    othek
    0.54
    idak
    0.53
    consin
    0.52
     SwiftUI
    0.52
     {}".
    0.51
    Activations Density 0.025%

    No Known Activations