Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 7920
    Prev
    Next
    INDEX
    Explanations

    technical terms and processes related to Fourier transformations and signal processing

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

    Fourier transform and frequency domain

    np_acts-logits-general · gemini-2.0-flash

    Fourier transform and frequency analysis

    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-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 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
    obut
    -0.56
    InjectMocks
    -0.55
     Peyton
    -0.54
    potranspiration
    -0.53
     Lari
    -0.52
    Varint
    -0.51
     BUB
    -0.50
    izmu
    -0.50
    uxta
    -0.49
     Chamb
    -0.49
    POSITIVE LOGITS
     Fourier
    1.06
     frequency
    0.96
     spectral
    0.86
    Fourier
    0.86
     frequencies
    0.84
     Frequency
    0.84
     spectrum
    0.81
     band
    0.79
    frequency
    0.77
     FFT
    0.77
    Activations Density 0.984%

    No Known Activations