© 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. 19443
    Prev
    Next
    INDEX
    Explanations

    verbs indicating physical or metaphorical movement away from a certain state or place

    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
    125
    +0.12
    0.5%
    724
    +0.10
    0.5%
    123
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    125
    +0.12
    0.05
    553
    +0.10
    0.04
    761
    +0.10
    0.04
    Negative Logits
    cest
    -0.73
    cius
    -0.69
    gow
    -0.66
    istani
    -0.61
    ourt
    -0.61
     Administ
    -0.60
     Founding
    -0.59
    ctions
    -0.58
     Tennis
    -0.57
    İĭ
    -0.56
    POSITIVE LOGITS
    wards
    0.95
    amorph
    0.92
     nicely
    0.89
     unexpectedly
    0.86
     smoothly
    0.85
     explos
    0.85
     abruptly
    0.84
     unpredict
    0.82
     spontaneously
    0.77
     exponentially
    0.77
    Activations Density 0.238%

    No Known Activations