© 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 3072
    9. 8-RES_FS3072-JB
    11. 198
    Prev
    Next
    INDEX
    Explanations

    instances where someone is the subject of actions or statements involving their actions by 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_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    391
    +0.10
    0.4%
    72
    +0.10
    0.4%
    155
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    718
    +0.10
    0.08
    671
    +0.10
    0.05
    762
    +0.09
    0.07
    Negative Logits
    idon
    -0.81
    atu
    -0.75
    ijing
    -0.68
    omas
    -0.65
    yrinth
    -0.65
    abre
    -0.65
    tower
    -0.64
    multiple
    -0.63
     manifold
    -0.62
    values
    -0.62
    POSITIVE LOGITS
     bothered
    1.00
     mention
    0.92
     remotely
    0.92
     blink
    0.91
     bother
    0.89
     bothering
    0.88
     ounce
    0.87
     scratch
    0.85
    touch
    0.80
     glance
    0.79
    Activations Density 0.917%

    No Known Activations