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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 10155
    Prev
    Next
    INDEX
    Explanations

    references to icons or iconic figures

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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%
    101
    +0.12
    0.6%
    571
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.13
    0.03
    701
    +0.12
    0.02
    326
    +0.11
    0.02
    Negative Logits
    gaard
    -0.67
    nda
    -0.65
    abouts
    -0.65
    ingly
    -0.64
    EEE
    -0.64
    ndum
    -0.61
    ood
    -0.61
     Thoughts
    -0.61
    interrupted
    -0.60
    THER
    -0.59
    POSITIVE LOGITS
    ocl
    1.58
    nect
    1.16
    ically
    1.14
    icity
    1.12
    ography
    1.09
    ographic
    1.06
    ographically
    1.01
    icon
    0.97
    stones
    0.92
    ologies
    0.91
    Activations Density 0.100%

    No Known Activations