© 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 49152
    9. 8-RES_FS49152-JB
    11. 43063
    Prev
    Next
    INDEX
    Explanations

    the word "and" used frequently in the context of discussions or lists

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    66
    +0.11
    0.5%
    393
    +0.10
    0.5%
    652
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    445
    +0.11
    0.01
    666
    +0.10
    0.01
    746
    +0.09
    0.01
    Negative Logits
    HUD
    -0.71
    catentry
    -0.71
    cess
    -0.66
    usk
    -0.66
    imon
    -0.65
    quet
    -0.64
    wake
    -0.61
    ROR
    -0.60
     cornerstone
    -0.60
    STEM
    -0.59
    POSITIVE LOGITS
     blindly
    0.80
     guess
    0.77
     indulge
    0.77
     purchase
    0.76
     install
    0.74
     let
    0.74
     extermin
    0.72
     marry
    0.72
     reprint
    0.71
     invade
    0.71
    Activations Density 0.014%

    No Known Activations