© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 4259
    Prev
    Next
    INDEX
    Explanations

    mentions of specific locations and settings, particularly involving people and events occurring in those places

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    593
    +0.11
    0.5%
    459
    +0.11
    0.5%
    182
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    764
    +0.11
    0.03
    526
    +0.11
    0.01
    79
    +0.10
    0.03
    Negative Logits
    iom
    -0.62
    Ĥİ
    -0.61
    inez
    -0.59
    lla
    -0.59
    +.
    -0.59
    ONSORED
    -0.58
    ²¾
    -0.57
    ela
    -0.57
    cially
    -0.57
    qua
    -0.57
    POSITIVE LOGITS
     seiz
    0.95
     chairs
    0.90
     bench
    0.84
     throne
    0.82
     benches
    0.82
     cub
    0.81
     chair
    0.81
    chair
    0.80
    isode
    0.77
     folding
    0.77
    Activations Density 0.133%

    No Known Activations