© 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. 12346
    Prev
    Next
    INDEX
    Explanations

    information related to database management and structured data operations

    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
    640
    +0.24
    1.1%
    481
    +0.14
    0.7%
    6
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.24
    0.14
    640
    +0.14
    0.11
    255
    +0.13
    0.10
    Negative Logits
     imperson
    -0.70
     blister
    -0.69
     inund
    -0.68
     unseen
    -0.68
     disemb
    -0.67
     thrill
    -0.66
     deadly
    -0.66
     savage
    -0.66
    oaded
    -0.66
     cherished
    -0.65
    POSITIVE LOGITS
     Lastly
    1.86
     Anyway
    1.77
     Furthermore
    1.57
     Additionally
    1.55
     Regardless
    1.54
     Nonetheless
    1.54
     Nevertheless
    1.53
     Finally
    1.52
     However
    1.50
     Secondly
    1.50
    Activations Density 0.441%

    No Known Activations