© 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. 17034
    Prev
    Next
    INDEX
    Explanations

    news articles talking about business developments or corporate activities, especially related to construction or expansion plans

    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
    +0.20
    1.0%
    373
    +0.18
    0.9%
    481
    +0.17
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.20
    0.07
    160
    +0.18
    0.07
    442
    +0.17
    0.06
    Negative Logits
     deduct
    -0.80
     dece
    -0.74
     triv
    -0.71
     monet
    -0.69
     favor
    -0.68
     crawling
    -0.68
     rg
    -0.67
     beh
    -0.66
     aggreg
    -0.65
     synerg
    -0.64
    POSITIVE LOGITS
    Published
    1.42
    Updated
    1.33
    Former
    1.24
    Australia
    1.18
    Australian
    1.17
    Britain
    1.15
    Ireland
    1.14
    Actor
    1.14
    TOR
    1.13
    PORT
    1.13
    Activations Density 0.145%

    No Known Activations