© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 49152
    9. 8-RES_FS49152-JB
    11. 43040
    Prev
    Next
    INDEX
    Explanations

    phrases indicating ownership or management of entities

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    464
    +0.11
    0.5%
    117
    +0.09
    0.4%
    197
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    197
    +0.11
    0.02
    108
    +0.09
    0.02
    498
    +0.09
    0.02
    Negative Logits
    pler
    -0.78
    izophren
    -0.78
    zh
    -0.76
    truth
    -0.75
    izoph
    -0.74
    ivable
    -0.70
    secut
    -0.69
    olithic
    -0.69
    ressing
    -0.68
    gart
    -0.68
    POSITIVE LOGITS
     subsidiaries
    0.87
     its
    0.75
     campuses
    0.74
     subcontract
    0.72
     offices
    0.72
     worldwide
    0.71
     facilities
    0.70
     Brach
    0.69
     Worldwide
    0.69
     partnerships
    0.68
    Activations Density 0.084%

    No Known Activations