© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 12843
    Prev
    Next
    INDEX
    Explanations

    historical references and storytelling elements

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    261
    +0.12
    0.5%
    259
    +0.10
    0.5%
    363
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    354
    +0.12
    0.05
    545
    +0.10
    0.03
    399
    +0.10
    0.03
    Negative Logits
    inventoryQuantity
    -1.07
     understatement
    -0.80
     redes
    -0.76
    vre
    -0.75
    phabet
    -0.73
    notations
    -0.71
    ãĤ´ãĥ³
    -0.70
    ãĤ¦ãĤ¹
    -0.69
     resilience
    -0.69
     improvement
    -0.68
    POSITIVE LOGITS
     Suddenly
    0.92
     suddenly
    0.91
    Suddenly
    0.85
    uddenly
    0.83
     overheard
    0.82
     casually
    0.81
     Suppose
    0.81
     backstage
    0.81
     randomly
    0.81
     approached
    0.79
    Activations Density 0.804%

    No Known Activations