© 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. 17099
    Prev
    Next
    INDEX
    Explanations

    words related to variety or options

    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
    362
    +0.12
    0.5%
    288
    +0.09
    0.4%
    751
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    658
    +0.12
    0.04
    751
    +0.09
    0.03
    569
    +0.09
    0.03
    Negative Logits
    Never
    -0.77
     âĨij
    -0.76
    ³³³³³³³³³³³³³³³³
    -0.73
    uble
    -0.72
    WARD
    -0.68
    NB
    -0.67
    never
    -0.66
    OIL
    -0.66
    âĢł
    -0.66
    ahead
    -0.66
    POSITIVE LOGITS
    iating
    1.77
     kinds
    1.55
    iates
    1.44
    iations
    1.35
     types
    1.35
    ials
    1.30
     versions
    1.26
     styles
    1.19
     combinations
    1.14
     interpretations
    1.14
    Activations Density 0.050%

    No Known Activations