© 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. 0-RES-JB
    11. 2064
    Prev
    Next
    INDEX
    Explanations

    text related to reputation, both positive and negative

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.0.hook_resid_post_16384_anthropic
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    HuggingFaceFW/fineweb
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    50
    +0.12
    0.8%
    528
    +0.04
    0.3%
    1870
    +0.03
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1385
    +0.12
    0.10
    1870
    +0.04
    0.02
    24
    +0.03
    0.07
    Negative Logits
    <bos>
    -2.13
    ⓧ
    -0.97
    /**
    -0.93
    
    -0.93
    <?
    -0.89
    /***
    -0.87
    /*
    -0.83
    <?
    -0.82
    ///**
    -0.78
    /*++
    -0.70
    POSITIVE LOGITS
     wien
    1.46
     maroc
    1.46
     affor
    1.40
     lele
    1.39
     meis
    1.36
     aen
    1.36
     bandung
    1.35
     véhic
    1.35
     stockholm
    1.34
     lidl
    1.30
    Activations Density 0.875%

    No Known Activations