Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 41311
    Prev
    Next
    INDEX
    Explanations

    specific years and dates associated with events or milestones

    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
    373
    +0.18
    0.8%
    430
    +0.12
    0.5%
    544
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    306
    +0.18
    0.06
    430
    +0.12
    0.05
    598
    +0.12
    0.04
    Negative Logits
    EStreamFrame
    -0.72
     ancest
    -0.71
    vec
    -0.64
    dden
    -0.63
     VIDEOS
    -0.61
     webcam
    -0.60
    nces
    -0.58
    kef
    -0.57
     sofa
    -0.57
    olor
    -0.56
    POSITIVE LOGITS
     onwards
    0.90
     onward
    0.85
    fml
    0.69
    -'
    0.63
    orpor
    0.62
    argo
    0.60
    ello
    0.59
    after
    0.59
     thereafter
    0.59
     amended
    0.57
    Activations Density 0.103%

    No Known Activations