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 24576
    5. 8-RES_FS24576-JB
    6. 21328
    Prev
    Next
    INDEX
    Explanations

    information related to sports drafts and player analysis

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    633
    +0.12
    0.5%
    17
    +0.11
    0.5%
    761
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    603
    +0.12
    0.05
    761
    +0.11
    0.04
    122
    +0.10
    0.04
    Negative Logits
     realise
    -0.79
     apologise
    -0.77
     Paradox
    -0.74
     realised
    -0.73
    pour
    -0.72
     apologised
    -0.72
     cease
    -0.72
     civilisation
    -0.72
    choes
    -0.72
    idays
    -0.71
    POSITIVE LOGITS
     athlet
    1.37
     athleticism
    1.33
     athletic
    1.19
     undrafted
    1.15
     versatility
    1.05
     scouts
    1.04
     NCAA
    1.04
    College
    1.03
    Height
    1.00
     collegiate
    0.97
    Activations Density 0.389%

    No Known Activations