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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 65k
    5. 20-GEMMASCOPE-RES-65K
    6. 1139
    Prev
    Next
    INDEX
    Explanations

    terms related to coding theory and error rates in data communication

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

    terms related to coding theory and its complexity in information processing

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

    error probability and secrecy

    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-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    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
    jvm
    -0.46
    cipital
    -0.44
    laf
    -0.43
    vac
    -0.43
     वै
    -0.42
    maakt
    -0.42
     prestige
    -0.41
     Culture
    -0.40
     Ke
    -0.40
    islav
    -0.40
    POSITIVE LOGITS
     signal
    0.78
     CreateTagHelper
    0.73
     transmitter
    0.70
     channel
    0.70
     receiver
    0.69
     MessageBoxIcon
    0.69
     receivers
    0.65
     Channel
    0.64
    signal
    0.64
     trasm
    0.63
    Activations Density 0.450%

    No Known Activations