© 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. 6984
    Prev
    Next
    INDEX
    Explanations

    phrases related to addressing inconsistency or putting something behind to move forward

    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
    458
    +0.12
    0.6%
    650
    +0.11
    0.5%
    408
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    526
    +0.12
    0.00
    458
    +0.11
    0.04
    764
    +0.11
    0.03
    Negative Logits
    DragonMagazine
    -0.88
     Cosponsors
    -0.75
    atri
    -0.74
    ARCH
    -0.63
    racuse
    -0.62
    anian
    -0.62
     Rooms
    -0.62
    emonium
    -0.60
    ynski
    -0.59
    interstitial
    -0.59
    POSITIVE LOGITS
     shoulders
    0.71
     unfounded
    0.65
    ¯
    0.65
     equation
    0.64
     sails
    0.64
     manageable
    0.64
     inhib
    0.63
     tires
    0.62
     easier
    0.61
     simpl
    0.60
    Activations Density 0.243%

    No Known Activations