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. MLP Out - 65k
    5. 4-GEMMASCOPE-MLP-65K
    6. 42
    Prev
    Next
    INDEX
    Explanations

    references to probing or investigation techniques

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 4-gemmascope-mlp-65k
    Configuration
    google/gemma-scope-2b-pt-mlp/layer_4/width_65k/average_l0_66
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.4.hook_mlp_out
    Hook Layer
    4
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     probe
    -2.36
     Probe
    -2.13
     probes
    -2.09
    probe
    -2.08
    Probe
    -2.05
     Probes
    -1.77
     probing
    -1.59
    PROBE
    -1.59
     probed
    -1.53
     sonda
    -1.05
    POSITIVE LOGITS
    moiselle
    0.56
    ỳnh
    0.55
    currentThread
    0.54
    iosync
    0.51
    Paginator
    0.50
    oplayer
    0.49
    脚注の使い方
    0.49
    ücksich
    0.49
     dangereux
    0.49
    RTLU
    0.49
    Activations Density 0.016%

    No Known Activations