© 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 24576
    9. 8-RES_FS24576-JB
    11. 24276
    Prev
    Next
    INDEX
    Explanations

    phrases and words related to reduction or minimizing

    oai_token-act-pair · gpt-4o-miniTriggered 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
    421
    +0.10
    0.5%
    624
    +0.10
    0.4%
    71
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    184
    +0.10
    0.03
    766
    +0.10
    0.03
    421
    +0.10
    0.03
    Negative Logits
    finished
    -0.65
    atom
    -0.64
     Remastered
    -0.64
    cart
    -0.63
    thing
    -0.62
    spr
    -0.62
    Reviewed
    -0.61
     Sina
    -0.61
    Poké
    -0.60
    shake
    -0.60
    POSITIVE LOGITS
     reliance
    0.93
     inhib
    0.88
     visibility
    0.87
     friction
    0.86
     emissions
    0.85
     dependence
    0.84
     effectiveness
    0.84
     workload
    0.83
     incidence
    0.82
     susceptibility
    0.81
    Activations Density 0.055%

    No Known Activations