© 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 12288
    9. 8-RES_FS12288-JB
    11. 579
    Prev
    Next
    INDEX
    Explanations

    conversations and activities related to social interactions and personal relationships

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    517
    +0.10
    0.5%
    393
    +0.09
    0.4%
    644
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    43
    +0.10
    0.07
    546
    +0.09
    0.07
    517
    +0.09
    0.07
    Negative Logits
    ¤
    -0.81
    encies
    -0.79
    gression
    -0.77
    isi
    -0.74
    UF
    -0.74
    end
    -0.74
    itionally
    -0.73
    eus
    -0.72
    GF
    -0.71
    olo
    -0.70
    POSITIVE LOGITS
     surrounded
    1.23
     flanked
    1.23
     clutching
    1.07
     awaiting
    1.06
     grinning
    1.02
     gazing
    1.02
     whispering
    1.01
     overlooking
    1.01
     smelling
    1.01
     ready
    1.00
    Activations Density 0.567%

    No Known Activations