© 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. 24011
    Prev
    Next
    INDEX
    Explanations

    instances of the word "In" indicating the start of explanations or contexts

    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
    640
    +0.12
    0.6%
    123
    +0.12
    0.5%
    91
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    123
    +0.12
    0.07
    91
    +0.12
    0.07
    257
    +0.11
    0.06
    Negative Logits
     looms
    -0.70
     unemploy
    -0.66
    76561
    -0.63
     disgr
    -0.61
     blasted
    -0.60
     smack
    -0.60
     entertained
    -0.60
     lodged
    -0.59
     nailed
    -0.59
    resh
    -0.59
    POSITIVE LOGITS
     addition
    1.44
     contrast
    1.33
    juries
    1.26
    herent
    1.24
    jured
    1.20
    ventory
    1.19
     spite
    1.17
    clusion
    1.16
    cluding
    1.15
    expensive
    1.14
    Activations Density 0.144%

    No Known Activations