© 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 24576
    9. 8-RES_FS24576-JB
    11. 8731
    Prev
    Next
    INDEX
    Explanations

    streaming-related terms

    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
    178
    +0.14
    0.6%
    289
    +0.10
    0.5%
    459
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    178
    +0.14
    0.02
    361
    +0.10
    0.01
    439
    +0.10
    0.01
    Negative Logits
     reluct
    -0.77
    ailability
    -0.67
     misunderstanding
    -0.67
    姫
    -0.65
     misunder
    -0.64
    cules
    -0.64
     Mans
    -0.61
     Palestin
    -0.60
    Downloadha
    -0.58
    nces
    -0.58
    POSITIVE LOGITS
    lining
    1.41
    lined
    1.14
    liner
    1.07
    line
    1.03
    lines
    0.99
    ers
    0.93
    stream
    0.91
    liners
    0.88
    atcher
    0.87
    flow
    0.84
    Activations Density 0.021%

    No Known Activations