© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 20609
    Prev
    Next
    INDEX
    Explanations

    mentions or references to activities and events related to running or a run in various contexts

    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
    627
    +0.10
    0.4%
    607
    +0.10
    0.4%
    270
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    422
    +0.10
    0.03
    540
    +0.10
    0.03
    607
    +0.10
    0.02
    Negative Logits
     Hots
    -0.71
    ĵĺ
    -0.67
    alam
    -0.64
     Males
    -0.61
     Virtue
    -0.60
     Photographer
    -0.60
     Ink
    -0.60
    ĨĴ
    -0.60
     defic
    -0.58
     defe
    -0.58
    POSITIVE LOGITS
    aways
    1.29
    gs
    1.21
    nings
    1.10
    away
    1.09
    escape
    1.09
    ners
    1.08
    ways
    1.06
    swick
    1.06
    dy
    0.99
    gression
    0.96
    Activations Density 0.037%

    No Known Activations