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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 457
    Prev
    Next
    INDEX
    Explanations

    references to characters or entities from superhero movies

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    515
    +0.13
    0.6%
    247
    +0.10
    0.5%
    314
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    515
    +0.13
    0.06
    247
    +0.10
    0.05
    232
    +0.10
    0.05
    Negative Logits
    PsyNetMessage
    -0.79
    RIPT
    -0.71
    TING
    -0.66
    IFT
    -0.65
    èĥ
    -0.63
    åī
    -0.62
    ="#
    -0.62
    âĨij
    -0.61
    ython
    -0.60
    REC
    -0.59
    POSITIVE LOGITS
    uscript
    1.44
    hattan
    1.38
    hunt
    1.26
    ifest
    1.25
    fred
    1.23
    oeuv
    1.22
    hood
    1.22
    nered
    1.21
    osphere
    1.13
    WithNo
    1.10
    Activations Density 0.597%

    No Known Activations