© 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. 1240
    Prev
    Next
    INDEX
    Explanations

    references to technology and digital markets

    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
    36
    +0.25
    1.2%
    393
    +0.17
    0.8%
    640
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.25
    0.14
    436
    +0.17
    0.06
    393
    +0.10
    0.06
    Negative Logits
    Versions
    -0.93
    units
    -0.79
    Edit
    -0.76
    notes
    -0.76
    effects
    -0.75
    alks
    -0.72
    agues
    -0.71
    uin
    -0.71
    Links
    -0.70
    onduct
    -0.70
    POSITIVE LOGITS
     lot
    1.18
     reminder
    1.12
     fascinating
    1.11
     huge
    1.10
     bit
    1.04
     daunting
    1.04
     coincidence
    1.03
     continuation
    1.01
     nice
    1.00
     HUGE
    0.99
    Activations Density 0.290%

    No Known Activations