© 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. 45574
    Prev
    Next
    INDEX
    Explanations

    the word "all" in various contexts

    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
    718
    +0.11
    0.5%
    416
    +0.10
    0.5%
    619
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    416
    +0.11
    0.04
    619
    +0.10
    0.04
    534
    +0.10
    0.03
    Negative Logits
    aminer
    -0.61
    liest
    -0.60
    IDS
    -0.59
    bal
    -0.58
    illin
    -0.57
    hift
    -0.57
    oute
    -0.57
    ahime
    -0.57
    hap
    -0.56
    stood
    -0.55
    POSITIVE LOGITS
    ocating
    1.13
    igator
    1.13
    uding
    1.12
    igators
    1.00
    usion
    0.99
    udes
    0.98
    usions
    0.95
    together
    0.93
    ocated
    0.93
    uring
    0.91
    Activations Density 0.042%

    No Known Activations