© 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 49152
    9. 8-RES_FS49152-JB
    11. 26641
    Prev
    Next
    INDEX
    Explanations

    phrases indicating quantities or measurements in a descriptive context

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    +0.14
    0.6%
    393
    +0.11
    0.5%
    53
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.14
    0.03
    393
    +0.11
    0.04
    289
    +0.10
    0.01
    Negative Logits
    Reviewer
    -0.76
    ummer
    -0.75
     Anxiety
    -0.75
    ibling
    -0.73
     Ratings
    -0.72
     Intervention
    -0.71
    istar
    -0.71
     Teachers
    -0.70
    taker
    -0.69
     Week
    -0.69
    POSITIVE LOGITS
     clay
    1.21
     limestone
    1.20
     ceramic
    1.19
     asphalt
    1.16
     acrylic
    1.16
     pulp
    1.15
     marble
    1.14
     resin
    1.14
     gravel
    1.13
     linen
    1.12
    Activations Density 0.128%

    No Known Activations