© 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. 20175
    Prev
    Next
    INDEX
    Explanations

    phrases related to time and actions

    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
    373
    +0.13
    0.6%
    589
    +0.10
    0.4%
    350
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    589
    +0.13
    0.02
    526
    +0.10
    0.00
    323
    +0.09
    0.02
    Negative Logits
    iple
    -0.81
    ories
    -0.81
    /-
    -0.75
    orously
    -0.69
    anship
    -0.69
    imens
    -0.68
    vari
    -0.68
    rams
    -0.67
    mt
    -0.66
    ivities
    -0.66
    POSITIVE LOGITS
     cue
    0.72
     doorstep
    0.68
     Shooter
    0.60
     Gomez
    0.60
     Springer
    0.59
    othal
    0.56
    velength
    0.56
    soDeliveryDate
    0.56
     borders
    0.56
     Grimes
    0.56
    Activations Density 0.057%

    No Known Activations