© 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. 20402
    Prev
    Next
    INDEX
    Explanations

    actions or activities related to personal experiences or interactions with others

    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
    183
    +0.14
    0.7%
    566
    +0.14
    0.6%
    552
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    566
    +0.14
    0.06
    183
    +0.14
    0.06
    552
    +0.12
    0.05
    Negative Logits
     precursor
    -0.69
     sidelines
    -0.67
     Apart
    -0.65
    earable
    -0.62
    urden
    -0.60
    atically
    -0.60
     sweeps
    -0.59
    bies
    -0.59
     Gutenberg
    -0.59
     Georgian
    -0.58
    POSITIVE LOGITS
     myself
    1.05
     personally
    1.04
     been
    1.03
    been
    0.96
     heard
    0.93
     seen
    0.91
     noticed
    0.87
     subscribed
    0.87
     gotten
    0.85
     never
    0.85
    Activations Density 0.138%

    No Known Activations