© 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. 6554
    Prev
    Next
    INDEX
    Explanations

    technical terms related to numerical data and analysis

    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
    488
    +0.09
    0.4%
    90
    +0.09
    0.4%
    145
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.09
    0.02
    145
    +0.09
    0.02
    488
    +0.09
    0.02
    Negative Logits
     msec
    -0.59
    DonaldTrump
    -0.59
     intentionally
    -0.58
     vacuum
    -0.56
     derog
    -0.55
    regon
    -0.55
    emort
    -0.54
     Hort
    -0.53
     contrace
    -0.52
     abbrevi
    -0.52
    POSITIVE LOGITS
    XM
    0.96
    ZX
    0.95
    dq
    0.92
    qq
    0.90
    CN
    0.89
    FK
    0.88
    Uk
    0.86
    wr
    0.86
    Fu
    0.86
    zx
    0.86
    Activations Density 0.032%

    No Known Activations