© 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 49152
    9. 8-RES_FS49152-JB
    11. 41583
    Prev
    Next
    INDEX
    Explanations

    URLs and links

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    640
    +0.12
    0.6%
    373
    +0.11
    0.5%
    313
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    107
    +0.12
    0.02
    313
    +0.11
    0.02
    412
    +0.10
    0.02
    Negative Logits
     interior
    -0.73
     conclud
    -0.71
     architecture
    -0.70
     inaccur
    -0.69
     retali
    -0.69
     decor
    -0.69
    ailability
    -0.67
     inward
    -0.66
     ��������
    -0.65
     signage
    -0.64
    POSITIVE LOGITS
    DragonMagazine
    1.04
    dp
    1.02
    dq
    1.02
    zx
    1.01
    76561
    1.00
    ubuntu
    0.99
    projects
    0.99
    dx
    0.96
    download
    0.94
    gd
    0.93
    Activations Density 0.011%

    No Known Activations