Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 1536
    5. 8-RES_FS1536-JB
    6. 734
    Prev
    Next
    INDEX
    Explanations

    statistics and performance metrics related to sports, especially baseball

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    668
    +0.12
    0.6%
    238
    +0.11
    0.5%
    383
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    238
    +0.12
    0.12
    412
    +0.11
    0.10
    260
    +0.10
    0.11
    Negative Logits
     Shant
    -0.76
    forums
    -0.75
    cul
    -0.71
    yg
    -0.70
    sylv
    -0.65
    igm
    -0.61
     Krish
    -0.61
     eth
    -0.60
    morrow
    -0.60
    actionDate
    -0.59
    POSITIVE LOGITS
     assists
    1.07
     postseason
    1.04
     batted
    1.04
     touchdowns
    1.01
     averaging
    1.00
     Rookie
    1.00
     scoring
    1.00
     interceptions
    0.99
     outings
    0.98
     streak
    0.97
    Activations Density 2.527%

    No Known Activations