© 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 · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 5042
    Prev
    Next
    INDEX
    Explanations

    references to news stories

    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
    160
    +0.12
    0.6%
    195
    +0.11
    0.5%
    6
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.12
    0.04
    634
    +0.11
    0.04
    247
    +0.11
    0.03
    Negative Logits
    ategory
    -0.84
    aez
    -0.78
    ignt
    -0.76
    ynski
    -0.72
    entimes
    -0.72
    amily
    -0.72
    anyon
    -0.71
    achev
    -0.70
    owered
    -0.70
    ojure
    -0.70
    POSITIVE LOGITS
    telling
    1.15
    te
    0.95
     arc
    0.89
     reprinted
    0.87
    book
    0.86
    board
    0.83
     arcs
    0.82
    icle
    0.82
    books
    0.81
     involving
    0.81
    Activations Density 0.047%

    No Known Activations