© 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. 752
    Prev
    Next
    INDEX
    Explanations

    references to legal or military matters, such as detainee treatment and military mishaps

    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
    447
    +0.18
    0.8%
    5
    +0.10
    0.4%
    276
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    8
    +0.18
    0.10
    608
    +0.10
    0.12
    171
    +0.09
    0.08
    Negative Logits
    Tonight
    -0.63
    sis
    -0.62
    orum
    -0.59
    craft
    -0.58
    click
    -0.58
    itas
    -0.57
    CPU
    -0.57
    perfect
    -0.56
     !=
    -0.55
     >=
    -0.55
    POSITIVE LOGITS
    cellaneous
    0.76
    interstitial
    0.72
     secondly
    0.71
     additionally
    0.71
    senal
    0.69
     newcom
    0.68
     indirectly
    0.68
    merce
    0.68
     optionally
    0.68
     obligatory
    0.65
    Activations Density 2.999%

    No Known Activations