© 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. 41544
    Prev
    Next
    INDEX
    Explanations

    phrases related to health and the impacts of chemicals or environmental factors

    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
    640
    +0.21
    1.0%
    135
    +0.13
    0.6%
    255
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.21
    0.08
    640
    +0.13
    0.07
    255
    +0.12
    0.06
    Negative Logits
    mble
    -0.83
    itage
    -0.82
     footing
    -0.81
     tenure
    -0.76
     delegated
    -0.75
     flagship
    -0.74
     categ
    -0.74
     planned
    -0.74
     enterprise
    -0.72
    edom
    -0.72
    POSITIVE LOGITS
     Symptoms
    1.84
     Researchers
    1.35
     However
    1.35
     Diseases
    1.31
     Therefore
    1.30
     Fortunately
    1.29
     Additionally
    1.28
     Scientists
    1.28
     Depending
    1.28
     Luckily
    1.26
    Activations Density 0.242%

    No Known Activations