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 3072
    5. 8-RES_FS3072-JB
    6. 410
    Prev
    Next
    INDEX
    Explanations

    phrases related to focus and attention

    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
    608
    +0.11
    0.5%
    587
    +0.10
    0.5%
    589
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.11
    0.13
    669
    +0.10
    0.08
    587
    +0.10
    0.09
    Negative Logits
    DragonMagazine
    -1.06
    zynski
    -0.88
    lich
    -0.72
    omsky
    -0.72
    soever
    -0.72
    NS
    -0.70
     CLSID
    -0.67
    opter
    -0.67
    named
    -0.67
    listed
    -0.67
    POSITIVE LOGITS
     preserving
    1.20
     maximizing
    1.14
     improving
    1.12
     maintaining
    1.08
     optimizing
    1.06
     achieving
    1.05
     solving
    1.00
     ensuring
    0.96
     aesthetics
    0.95
     protecting
    0.94
    Activations Density 1.324%

    No Known Activations