© 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. 892
    Prev
    Next
    INDEX
    Explanations

    phrases related to physical actions or movements

    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
    603
    +0.13
    0.6%
    168
    +0.13
    0.6%
    351
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    168
    +0.13
    0.03
    603
    +0.13
    0.03
    125
    +0.10
    0.02
    Negative Logits
    cest
    -0.84
    aga
    -0.65
    imar
    -0.63
    MAT
    -0.63
    ################################
    -0.61
    sen
    -0.61
     coh
    -0.60
    BIT
    -0.60
    ided
    -0.59
    imb
    -0.59
    POSITIVE LOGITS
     unexpectedly
    0.74
    stairs
    0.73
     prominently
    0.71
    doors
    0.71
     anywhere
    0.70
     randomly
    0.70
    nces
    0.70
     alongside
    0.68
    number
    0.68
     smelling
    0.68
    Activations Density 0.095%

    No Known Activations