© 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 12288
    9. 8-RES_FS12288-JB
    11. 1525
    Prev
    Next
    INDEX
    Explanations

    verbs ending in 'ing' related to actions or behaviors

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    447
    +0.16
    0.7%
    90
    +0.10
    0.4%
    481
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    761
    +0.16
    0.05
    62
    +0.10
    0.04
    271
    +0.10
    0.04
    Negative Logits
    ère
    -0.75
    see
    -0.67
    ersen
    -0.63
    si
    -0.62
    Dim
    -0.61
    ,—
    -0.59
    ses
    -0.58
    binding
    -0.58
    word
    -0.57
    sg
    -0.57
    POSITIVE LOGITS
     oneself
    0.96
     Yourself
    0.78
     yourself
    0.74
    inventoryQuantity
    0.69
    redients
    0.65
     Basics
    0.65
     truthful
    0.62
    HAM
    0.61
    ealous
    0.61
     Ahead
    0.61
    Activations Density 0.354%

    No Known Activations