© 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 24576
    9. 8-RES_FS24576-JB
    11. 24228
    Prev
    Next
    INDEX
    Explanations

    phrases indicating first-time achievements or significant milestones

    oai_token-act-pair · gpt-4o-miniTriggered 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
    694
    +0.11
    0.5%
    383
    +0.10
    0.5%
    140
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    366
    +0.11
    0.05
    339
    +0.10
    0.05
    162
    +0.10
    0.04
    Negative Logits
    bara
    -0.88
     Vish
    -0.73
    stairs
    -0.72
    halla
    -0.64
    rox
    -0.63
    blocks
    -0.63
    mund
    -0.63
    forts
    -0.63
    sometimes
    -0.62
    facts
    -0.62
    POSITIVE LOGITS
     foray
    1.45
     outing
    1.21
     birthday
    1.03
     appearance
    0.97
     incarnation
    0.96
     attempt
    0.96
     stint
    0.96
     cousin
    0.93
     playthrough
    0.90
     ever
    0.90
    Activations Density 0.098%

    No Known Activations