© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 22821
    Prev
    Next
    INDEX
    Explanations

    supportive interactions or references to assistance

    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
    288
    +0.13
    0.6%
    679
    +0.12
    0.6%
    540
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    540
    +0.13
    0.03
    417
    +0.12
    0.03
    225
    +0.12
    0.03
    Negative Logits
    tein
    -0.72
    apeake
    -0.64
    ãĥ£
    -0.64
    76561
    -0.61
    ucket
    -0.60
    kered
    -0.59
     pores
    -0.59
     sweat
    -0.58
     Hebdo
    -0.58
    sung
    -0.57
    POSITIVE LOGITS
    ament
    0.82
    ive
    0.80
    bands
    0.79
     Vector
    0.78
    Support
    0.76
    heses
    0.75
     mechanisms
    0.73
     providers
    0.72
     services
    0.72
     provider
    0.71
    Activations Density 0.041%

    No Known Activations