© 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. 10816
    Prev
    Next
    INDEX
    Explanations

    phrases related to events or announcements

    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
    373
    +0.21
    1.0%
    447
    +0.20
    0.9%
    185
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    346
    +0.21
    0.06
    668
    +0.20
    0.04
    244
    +0.10
    0.05
    Negative Logits
    imar
    -0.67
     denies
    -0.62
     depends
    -0.62
    NOW
    -0.62
    ino
    -0.60
     tomorrow
    -0.60
     loves
    -0.60
     doesnt
    -0.60
     ASAP
    -0.60
     faiths
    -0.59
    POSITIVE LOGITS
    FTWARE
    0.76
    Reviewer
    0.75
    USS
    0.74
    perty
    0.67
     announcing
    0.66
    schild
    0.66
    itled
    0.65
    ixty
    0.63
     unexpectedly
    0.63
     again
    0.63
    Activations Density 1.029%

    No Known Activations