© 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. 1797
    Prev
    Next
    INDEX
    Explanations

    capital `F` followed by a digit specifically 9

    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
    314
    +0.13
    0.6%
    63
    +0.11
    0.5%
    330
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    330
    +0.13
    0.06
    103
    +0.11
    0.06
    552
    +0.11
    0.05
    Negative Logits
     machine
    -0.70
     mater
    -0.68
     Hole
    -0.67
    vier
    -0.65
     Reborn
    -0.64
     Hades
    -0.62
     Patriarch
    -0.62
     Boone
    -0.62
     Pv
    -0.61
     Hera
    -0.60
    POSITIVE LOGITS
    ortun
    1.40
    ocused
    1.35
    ruits
    1.34
    avour
    1.34
    ortunate
    1.33
    ocusing
    1.33
    ertility
    1.26
    rozen
    1.24
    ierce
    1.23
    avorable
    1.22
    Activations Density 0.071%

    No Known Activations