© 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. 41674
    Prev
    Next
    INDEX
    Explanations

    features related to product specifications and technology

    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
    +0.15
    0.7%
    288
    +0.12
    0.5%
    678
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.15
    0.04
    288
    +0.12
    0.05
    110
    +0.11
    0.04
    Negative Logits
    ealous
    -0.95
    cffffcc
    -0.85
    know
    -0.83
    =-=-=-=-
    -0.83
    threatening
    -0.81
    #$#$
    -0.78
    foreseen
    -0.77
    blogs
    -0.76
    someone
    -0.75
    acca
    -0.75
    POSITIVE LOGITS
     polymer
    1.00
     solution
    0.99
     model
    0.98
     version
    0.94
     aluminum
    0.94
     printer
    0.91
     variant
    0.91
     edition
    0.90
     alloy
    0.85
     hybrid
    0.85
    Activations Density 0.156%

    No Known Activations