© 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. 8728
    Prev
    Next
    INDEX
    Explanations

    references to methods of communication and connection

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

    This neuron detects preposition phrases using "via" and identifies methods of transmission or communication.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    phrases and patterns describing communication methods or channels (such as "via email", "via contact form", "via school media").

    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
     vôtre
    -0.55
    essentiel
    -0.50
     nôtre
    -0.49
    grunns
    -0.48
     koruma
    -0.48
     honneur
    -0.48
     infancia
    -0.44
     Wahrheit
    -0.43
    Và
    -0.43
     juventud
    -0.42
    POSITIVE LOGITS
     means
    1.18
     Means
    0.97
    Means
    0.93
     MEANS
    0.91
    means
    0.86
     via
    0.82
     посред
    0.78
     methods
    0.73
    手段
    0.71
     through
    0.66
    Activations Density 0.697%

    No Known Activations