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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 2160
    Prev
    Next
    INDEX
    Explanations

    instances of the word "read" in various contexts

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

    the word "read"

    np_acts-logits-general · gemini-2.0-flash

    reading actions

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    帖最后由
    -0.96
    ConstraintMaker
    -0.84
     MonoBehaviour
    -0.80
     AppCompatTheme
    -0.77
    makeConstraints
    -0.77
    AddTagHelper
    -0.74
    XmlAccessType
    -0.73
    Билгалдахарш
    -0.72
    Jegyzetek
    -0.69
    IntoConstraints
    -0.68
    POSITIVE LOGITS
     aloud
    0.98
     reads
    0.96
     READ
    0.93
     Reader
    0.92
     read
    0.88
    justed
    0.88
     Readers
    0.87
    Читайте
    0.87
     Reads
    0.86
    reads
    0.85
    Activations Density 0.202%

    No Known Activations