© 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. 26710
    Prev
    Next
    INDEX
    Explanations

    words related to negative outcomes or distressing experiences

    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
    716
    +0.10
    0.5%
    494
    +0.10
    0.5%
    101
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.10
    0.02
    110
    +0.10
    0.01
    181
    +0.10
    0.01
    Negative Logits
    inoa
    -0.74
     Iw
    -0.74
    taboola
    -0.73
     Tsukuyomi
    -0.70
     Debor
    -0.68
    eva
    -0.68
     Ezek
    -0.67
     Niet
    -0.66
     Sonia
    -0.65
    âĵĺ
    -0.64
    POSITIVE LOGITS
    icultural
    0.81
    aday
    0.78
    quarters
    0.73
    angs
    0.70
    lights
    0.70
    abee
    0.66
    sters
    0.66
    dog
    0.66
    acan
    0.65
    ogeneous
    0.65
    Activations Density 0.033%

    No Known Activations