© 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. 8571
    Prev
    Next
    INDEX
    Explanations

    the word "complete" followed by a number

    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
    137
    +0.11
    0.5%
    649
    +0.10
    0.4%
    462
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    137
    +0.11
    0.02
    655
    +0.10
    0.02
    207
    +0.09
    0.02
    Negative Logits
    maid
    -0.91
    hao
    -0.78
    hops
    -0.75
    spr
    -0.71
    uay
    -0.69
    */(
    -0.66
     Zub
    -0.66
    adish
    -0.65
    anwhile
    -0.65
    UL
    -0.65
    POSITIVE LOGITS
     strangers
    1.01
    bred
    0.94
     stranger
    0.85
     immersion
    0.84
     disregard
    0.83
     lack
    0.83
     absence
    0.81
     beginners
    0.80
     fabrication
    0.79
     annihilation
    0.79
    Activations Density 0.026%

    No Known Activations