© 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. 37592
    Prev
    Next
    INDEX
    Explanations

    verbs and their various conjugations in relation to subjects that imply identification or state of being

    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
    55
    +0.15
    0.7%
    102
    +0.14
    0.7%
    208
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    102
    +0.15
    0.09
    55
    +0.14
    0.09
    566
    +0.12
    0.07
    Negative Logits
    ousy
    -0.70
     Ago
    -0.69
    anka
    -0.67
     Higher
    -0.64
    inav
    -0.63
     IMAGES
    -0.63
     Moons
    -0.62
     dislike
    -0.62
     Lack
    -0.62
     occurs
    -0.61
    POSITIVE LOGITS
     poised
    1.30
     willing
    1.06
     primed
    1.03
     able
    1.01
     hoping
    0.99
     likely
    0.99
     destined
    0.97
     unlikely
    0.97
     ready
    0.97
     obligated
    0.96
    Activations Density 0.242%

    No Known Activations