© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 38-GEMMASCOPE-RES-16K
    6. 8268
    Prev
    Next
    INDEX
    Explanations

    coding structures related to flow control, such as loops and conditional statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 38-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_38/width_16k/average_l0_128
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.38.hook_resid_post
    Hook Layer
    38
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    (
    -0.66
    (“
    -0.65
    (!__
    -0.63
    (=
    -0.59
    ſelf
    -0.56
    (~
    -0.55
    ($__
    -0.54
    ſelves
    -0.54
    (‘
    -0.52
    (<
    -0.52
    POSITIVE LOGITS
     (
    0.55
     TAMBÉM
    0.44
     AssemblyCompany
    0.41
    WaitGroup
    0.41
    fromnode
    0.36
     (;
    0.36
     AssemblyTitle
    0.35
     ((
    0.35
    ValueStyle
    0.35
     ()
    0.34
    Activations Density 0.233%

    No Known Activations