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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Initial Gemma-2B SAEs
    5. Gemma-2B
    7. Residual Stream
    9. 10-RES-JB
    11. 3517
    Prev
    Next
    INDEX
    Explanations

    terms related to performance evaluation and metrics

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.10.hook_resid_post_16384
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    chanind/openwebtext-gemma
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    Architecture
    standard
    Context Size
    1,024
    Dataset
    ctigges/openwebtext-gemma-1024-cl
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    50
    +0.26
    1.6%
    1870
    +0.15
    1.0%
    528
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    783
    +0.26
    0.02
    341
    +0.15
    0.02
    528
    +0.12
    0.02
    Negative Logits
    <bos>
    -3.02
    /***
    -0.74
    <?
    -0.68
    ⓧ
    -0.64
    
    -0.63
    addFlags
    -0.62
    ɵɵ
    -0.62
    <?
    -0.61
    #![
    -0.61
    /*!
    -0.61
    POSITIVE LOGITS
     jaya
    1.80
     bandung
    1.65
     Minang
    1.54
     lele
    1.53
     thut
    1.50
     perfon
    1.44
     jawa
    1.44
     aen
    1.43
     kasa
    1.42
     saar
    1.41
    Activations Density 0.033%

    No Known Activations