© 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. 35
    Prev
    Next
    INDEX
    Explanations

    mentions of numerical specifics or quantities in contexts such as reports, planned events, and data

    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
    426
    +0.11
    0.5%
    667
    +0.10
    0.4%
    116
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    23
    +0.11
    0.13
    426
    +0.10
    0.12
    107
    +0.09
    0.11
    Negative Logits
     shroud
    -0.77
     shack
    -0.74
     Downs
    -0.70
     dispers
    -0.69
     closest
    -0.69
     assemb
    -0.69
     touches
    -0.68
     clad
    -0.68
     segreg
    -0.68
     nude
    -0.67
    POSITIVE LOGITS
    ľ
    1.14
    º
    1.14
    ¬
    1.11
    £
    1.10
    ¡
    1.04
    «
    1.04
    Ĵ
    1.01
    ®
    1.01
    ¹
    1.00
    Ĥ
    1.00
    Activations Density 0.463%

    No Known Activations