© 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. 1195
    Prev
    Next
    INDEX
    Explanations

    recipes and instructions on food preparation

    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
    124
    +0.11
    0.5%
    68
    +0.11
    0.5%
    537
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    124
    +0.11
    0.06
    537
    +0.11
    0.06
    680
    +0.11
    0.04
    Negative Logits
     Arabian
    -0.84
     Dickinson
    -0.77
     moder
    -0.66
     wedd
    -0.65
     parade
    -0.65
     opener
    -0.65
     pause
    -0.64
     sponsor
    -0.64
     dive
    -0.63
     antiv
    -0.63
    POSITIVE LOGITS
    years
    1.19
    sized
    1.13
    based
    1.12
    eyed
    1.10
    year
    1.08
    style
    1.06
    months
    1.06
    hour
    1.05
    digit
    1.04
    dimensional
    1.03
    Activations Density 0.135%

    No Known Activations