© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 9526
    Prev
    Next
    INDEX
    Explanations

    references to legal or regulatory matters

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    304
    +0.10
    0.5%
    427
    +0.10
    0.4%
    631
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    720
    +0.10
    0.07
    765
    +0.10
    0.07
    43
    +0.10
    0.05
    Negative Logits
    quished
    -1.00
    ukong
    -0.77
    hai
    -0.74
    izens
    -0.70
    ificent
    -0.69
    kered
    -0.69
    omo
    -0.68
     earthqu
    -0.65
    undai
    -0.65
    zhou
    -0.64
    POSITIVE LOGITS
     rather
    1.39
     whereas
    1.36
    rather
    1.34
     instead
    1.06
    instead
    0.98
     not
    0.98
     namely
    0.91
     regardless
    0.90
     anyway
    0.87
     irrespective
    0.87
    Activations Density 0.698%

    No Known Activations