© 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. 24009
    Prev
    Next
    INDEX
    Explanations

    strong expressions of emotion or impactful statements

    oai_token-act-pair · gpt-4o-miniTriggered 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
    719
    +0.10
    0.5%
    426
    +0.10
    0.5%
    68
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    23
    +0.10
    0.08
    719
    +0.10
    0.08
    317
    +0.10
    0.08
    Negative Logits
     Belg
    -0.77
     handshake
    -0.71
     shack
    -0.71
     confinement
    -0.71
     buggy
    -0.70
     dispers
    -0.69
     playbook
    -0.69
     machine
    -0.69
     Faust
    -0.68
     platinum
    -0.67
    POSITIVE LOGITS
    ª
    1.30
    ¹
    1.27
    Ń
    1.21
    ¤
    1.20
    Į
    1.16
    į
    1.14
    £
    1.13
    º
    1.13
    ¬
    1.12
    ¡
    1.12
    Activations Density 0.101%

    No Known Activations