© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 31-GEMMASCOPE-RES-16K
    11. 5152
    Prev
    Next
    INDEX
    Explanations

    phrases indicating historical events and legal agreements

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

    commas that separate clauses or list items in formal or historical text.

    oai_token-act-pair · claude-4-5-haikuTriggered by @emiglarou
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 31-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_31/width_16k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.31.hook_resid_post
    Hook Layer
    31
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     propOrder
    -0.60
     المعيارى
    -0.59
     فريبيس
    -0.58
     oftentimes
    -0.57
    bootstrapcdn
    -0.57
     nakalista
    -0.54
     ThemeData
    -0.53
    Datuak
    -0.53
     Wikiseite
    -0.53
    IndentedString
    -0.52
    POSITIVE LOGITS
    DIPSETTING
    0.41
     being
    0.37
    klä
    0.36
    IUrlHelper
    0.36
    care
    0.35
    being
    0.35
    Being
    0.35
    as
    0.34
    clusal
    0.34
    pro
    0.34
    Activations Density 0.321%

    No Known Activations