© 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. 41151
    Prev
    Next
    INDEX
    Explanations

    references to a specific individual or personal identifier

    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
    641
    +0.11
    0.5%
    8
    +0.10
    0.4%
    588
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    299
    +0.11
    0.01
    588
    +0.10
    0.01
    641
    +0.10
    0.01
    Negative Logits
    stood
    -0.68
    stage
    -0.67
    managed
    -0.64
    lift
    -0.64
    locked
    -0.64
    ãĤ°
    -0.63
     tolerate
    -0.62
    ting
    -0.61
    chest
    -0.61
    lain
    -0.60
    POSITIVE LOGITS
    ences
    1.12
    emi
    0.98
    ppo
    0.94
    oglu
    0.92
    otti
    0.90
    encia
    0.89
    ardo
    0.89
    zona
    0.88
    pe
    0.88
    ère
    0.86
    Activations Density 0.006%

    No Known Activations