© 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
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 1216
    Prev
    Next
    INDEX
    Explanations

    phrases related to conflicts, confrontations, and intense situations

    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
    659
    +0.14
    0.6%
    561
    +0.11
    0.5%
    588
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.14
    0.04
    588
    +0.11
    0.04
    251
    +0.11
    0.03
    Negative Logits
     lapse
    -0.71
     remembrance
    -0.71
    elsen
    -0.64
     laure
    -0.63
     disarm
    -0.63
     contender
    -0.63
    vironment
    -0.62
     CPC
    -0.60
    nown
    -0.59
    heit
    -0.59
    POSITIVE LOGITS
    ulpt
    1.33
    rupulous
    1.28
    ratch
    1.27
    apers
    1.22
    ouring
    1.19
    oops
    1.19
    ribe
    1.19
    rawling
    1.18
    rawled
    1.18
    ammers
    1.17
    Activations Density 0.054%

    No Known Activations