© 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 49152
    9. 8-RES_FS49152-JB
    11. 45848
    Prev
    Next
    INDEX
    Explanations

    mentions of rarity and statistics related to data or items

    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
    373
    +0.15
    0.7%
    188
    +0.10
    0.5%
    277
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.15
    -0.01
    6
    +0.10
    0.02
    530
    +0.10
    0.01
    Negative Logits
     spons
    -0.71
    verages
    -0.66
     stret
    -0.63
    incinn
    -0.60
     flared
    -0.58
     nationally
    -0.55
     arter
    -0.55
    uala
    -0.54
    National
    -0.54
    vered
    -0.53
    POSITIVE LOGITS
     replies
    0.79
    remlin
    0.67
     WARN
    0.64
    language
    0.64
    reply
    0.63
     reply
    0.63
     IPM
    0.63
     bytes
    0.61
     UTC
    0.60
     @@
    0.60
    Activations Density 0.023%

    No Known Activations