© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 1309
    Prev
    Next
    INDEX
    Explanations

    questions in a written format

    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
    321
    +0.10
    0.4%
    24
    +0.10
    0.4%
    621
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    607
    +0.10
    0.02
    739
    +0.10
    0.02
    340
    +0.09
    0.02
    Negative Logits
    rawdownloadcloneembedreportprint
    -0.76
    ords
    -0.75
    apsed
    -0.69
    alach
    -0.67
    lag
    -0.64
     broch
    -0.62
    ographed
    -0.59
    olves
    -0.58
    estial
    -0.58
     Proud
    -0.58
    POSITIVE LOGITS
    why
    1.38
     WHY
    1.35
     why
    1.25
     Why
    1.16
    Why
    1.11
    how
    1.08
    what
    1.07
     whether
    1.03
    whether
    1.02
     WHERE
    1.00
    Activations Density 0.150%

    No Known Activations