© 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. 1317
    Prev
    Next
    INDEX
    Explanations

    details or descriptions related to products or items, such as measurements, materials used, pricing, and features

    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
    608
    +0.16
    0.8%
    291
    +0.10
    0.5%
    549
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.16
    0.06
    549
    +0.10
    0.03
    237
    +0.09
    0.03
    Negative Logits
    avior
    -0.75
     Advertisement
    -0.69
    mods
    -0.68
    llor
    -0.67
    AAAAAAAA
    -0.63
    identally
    -0.60
    Ê
    -0.59
    NSA
    -0.59
    necess
    -0.58
    cz
    -0.57
    POSITIVE LOGITS
    tones
    0.78
     proceedings
    0.75
    otos
    0.74
     humankind
    0.73
     accompany
    0.69
     mankind
    0.66
     counteract
    0.65
     srf
    0.65
     compensate
    0.65
    neath
    0.63
    Activations Density 0.151%

    No Known Activations