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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 9547
    Prev
    Next
    INDEX
    Explanations

    words related to discussions or debates

    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
    289
    +0.12
    0.5%
    569
    +0.12
    0.5%
    208
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    289
    +0.12
    0.01
    569
    +0.12
    0.02
    208
    +0.10
    0.02
    Negative Logits
    glers
    -0.72
    aukee
    -0.70
    mberg
    -0.68
    ahime
    -0.67
    akeru
    -0.64
    \<
    -0.64
     Holmes
    -0.60
    Interstitial
    -0.60
    giving
    -0.60
    gerald
    -0.59
    POSITIVE LOGITS
    ussions
    1.25
    ourses
    1.22
    ourse
    1.18
    ontent
    1.18
    overed
    1.13
    erning
    1.13
    ursive
    1.11
    rete
    1.09
    iple
    1.08
    ussion
    1.08
    Activations Density 0.059%

    No Known Activations