© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 12288
    5. 8-RES_FS12288-JB
    6. 1297
    Prev
    Next
    INDEX
    Explanations

    phrases indicating importance or necessity

    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
    197
    +0.13
    0.6%
    589
    +0.10
    0.5%
    170
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.13
    0.07
    589
    +0.10
    0.04
    197
    +0.10
    0.04
    Negative Logits
    DragonMagazine
    -0.89
    imilar
    -0.76
    scl
    -0.73
     Cosponsors
    -0.71
     borrowed
    -0.70
    keys
    -0.69
    balls
    -0.66
    onyms
    -0.65
    wic
    -0.65
    lights
    -0.65
    POSITIVE LOGITS
     determining
    1.22
     survival
    1.16
     achieving
    1.16
     unlocking
    1.12
     maintaining
    1.12
     ensuring
    1.10
     sustaining
    1.06
     preserving
    1.05
     understanding
    1.02
     combating
    1.01
    Activations Density 0.343%

    No Known Activations