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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 37189
    Prev
    Next
    INDEX
    Explanations

    issues related to violence, abuse, and societal injustices

    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
    481
    +0.22
    1.1%
    447
    +0.14
    0.7%
    277
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.22
    0.06
    762
    +0.14
    0.05
    481
    +0.11
    0.04
    Negative Logits
    UTC
    -0.71
    MET
    -0.70
    Export
    -0.70
    Ry
    -0.69
    details
    -0.67
    Cal
    -0.64
    Shares
    -0.63
    Pitt
    -0.62
    Fish
    -0.61
    SCP
    -0.61
    POSITIVE LOGITS
     trauma
    0.89
     syndrome
    0.86
    ukemia
    0.78
     Syndrome
    0.77
    agar
    0.73
     fame
    0.73
     illness
    0.71
     abroad
    0.69
     oppression
    0.69
     sickness
    0.68
    Activations Density 0.198%

    No Known Activations