© 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. 7796
    Prev
    Next
    INDEX
    Explanations

    names of characters or individuals being involved in some kind of conflict or drama

    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
    87
    +0.38
    1.9%
    447
    +0.19
    1.0%
    481
    +0.17
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.38
    0.08
    232
    +0.19
    0.06
    317
    +0.17
    0.05
    Negative Logits
     AAA
    -0.70
     Parables
    -0.61
    ramid
    -0.56
     ��������
    -0.56
    ailability
    -0.54
    interstitial
    -0.54
     Dise
    -0.52
    LEASE
    -0.52
    tumblr
    -0.52
     downgrade
    -0.51
    POSITIVE LOGITS
     remembers
    0.70
     agrees
    0.67
     commented
    0.65
     knew
    0.65
     himself
    0.65
     replied
    0.64
     insisted
    0.64
     wrote
    0.63
     nodded
    0.62
     admits
    0.61
    Activations Density 0.289%

    No Known Activations