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. 17-GEMMASCOPE-RES-65K
    6. 28486
    Prev
    Next
    INDEX
    Explanations

    references to the word "Bal" and its variations, indicating a focus on a specific name or term associated with "Bal."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.17.hook_resid_post
    Hook Layer
    17
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    enterOuterAlt
    -0.84
     کیلومتر
    -0.80
    LikeLike
    -0.75
    >-->
    -0.74
     INTERESAR
    -0.73
     USP
    -0.72
    );\
    -0.70
     resourceCulture
    -0.69
    DoubleQuotes
    -0.69
    RegressionTest
    -0.69
    POSITIVE LOGITS
     Balfour
    1.19
     Bal
    1.10
     BAL
    1.08
    Bal
    1.05
     Balzac
    1.02
     Balt
    0.93
     bal
    0.91
    BAL
    0.88
     Balk
    0.84
     Balboa
    0.84
    Activations Density 0.017%

    No Known Activations