© 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. 10456
    Prev
    Next
    INDEX
    Explanations

    technical words related to equipment and mechanisms, such as "monitor", "buttons", "wheel", and "jack"

    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
    420
    +0.13
    0.6%
    347
    +0.12
    0.5%
    234
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    55
    +0.13
    0.09
    566
    +0.12
    0.07
    347
    +0.10
    0.07
    Negative Logits
    ivals
    -0.92
    doms
    -0.86
     scenarios
    -0.78
     lifestyles
    -0.75
     Events
    -0.74
     timelines
    -0.73
    rosso
    -0.72
     Statements
    -0.72
     narratives
    -0.72
     prospects
    -0.71
    POSITIVE LOGITS
     worn
    1.30
     protr
    1.22
     malfunction
    1.20
     pierced
    1.09
     corro
    1.08
     adorned
    1.08
     cracked
    1.05
     thicker
    1.04
     jammed
    1.01
    equipped
    1.00
    Activations Density 0.936%

    No Known Activations