© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 6682
    Prev
    Next
    INDEX
    Explanations

    numbers, especially when related to events or statistics

    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
    491
    +0.08
    0.4%
    680
    +0.08
    0.4%
    314
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    223
    +0.08
    0.04
    689
    +0.08
    0.04
    693
    +0.08
    0.04
    Negative Logits
    ELL
    -0.59
    OA
    -0.52
    ORN
    -0.51
    ahime
    -0.51
    pex
    -0.51
     Ge
    -0.50
     buildup
    -0.50
    achi
    -0.50
     drafting
    -0.50
    iller
    -0.49
    POSITIVE LOGITS
     times
    1.02
    uously
    0.98
     nights
    0.89
    aciously
    0.89
    fold
    0.87
     mornings
    0.85
    times
    0.82
    til
    0.79
    heartedly
    0.78
     days
    0.76
    Activations Density 0.152%

    No Known Activations