© 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. 4673
    Prev
    Next
    INDEX
    Explanations

    instances of the word "by," particularly in contexts involving actions or events

    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
    +0.31
    1.3%
    752
    +0.10
    0.4%
    946
    +0.08
    0.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    752
    +0.31
    0.05
    506
    +0.10
    0.04
    946
    +0.08
    0.04
    Negative Logits
    <bos>
    -3.46
    -0.78
    protected
    -0.77
    -0.76
    IsRequired
    -0.76
    log
    -0.75
                    
    -0.75
    -0.74
    BeginContext
    -0.74
    -0.74
    POSITIVE LOGITS
     Juf
    2.30
     affor
    2.14
     Augu
    2.13
     increa
    2.05
     ecru
    2.01
     swarovski
    1.92
     stockholm
    1.91
     Khart
    1.90
     guarante
    1.90
     hairc
    1.87
    Activations Density 0.274%

    No Known Activations