© 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 49152
    9. 8-RES_FS49152-JB
    11. 45807
    Prev
    Next
    INDEX
    Explanations

    references to new appointments, formations, or developments in organizations or initiatives

    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
    91
    +0.11
    0.5%
    408
    +0.10
    0.4%
    101
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    91
    +0.11
    0.02
    408
    +0.10
    0.02
    101
    +0.10
    0.02
    Negative Logits
    omever
    -0.81
    imar
    -0.78
     teasp
    -0.77
    fax
    -0.76
    lua
    -0.74
    ©¶æ
    -0.72
    orously
    -0.70
    CLAIM
    -0.69
    rss
    -0.69
    claimer
    -0.69
    POSITIVE LOGITS
     additions
    0.76
     arrivals
    0.72
     challenger
    0.69
     Georgian
    0.67
     edition
    0.66
     arrival
    0.66
     hostilities
    0.66
     è£ıç
    0.65
     Antiqu
    0.65
     Supreme
    0.64
    Activations Density 0.041%

    No Known Activations