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-2-9B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 207
    Prev
    Next
    INDEX
    Explanations

    negative numerical values and their properties

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

    numerical comparisons and mathematical operations involving numbers.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
    RegressionTest
    -0.62
    出版年
    -0.58
    richTextPanel
    -0.54
    phosa
    -0.51
    flores
    -0.51
     Chwiliwch
    -0.49
    liceerd
    -0.49
    DebuggerNonUser
    -0.48
     BytesLike
    -0.48
    $_['
    -0.47
    POSITIVE LOGITS
     negative
    0.72
    negative
    0.62
    Negative
    0.60
     Negative
    0.60
     negativo
    0.60
     negativos
    0.54
     negativ
    0.54
     négatif
    0.53
    =−
    0.52
     negativas
    0.52
    Activations Density 0.719%

    No Known Activations