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 49152
    5. 8-RES_FS49152-JB
    6. 23820
    Prev
    Next
    INDEX
    Explanations

    mostly numerical references and statements related to player statistics and retirement in a sports 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
    640
    +0.11
    0.5%
    241
    +0.10
    0.4%
    630
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.11
    0.06
    6
    +0.10
    0.05
    630
    +0.10
    0.03
    Negative Logits
    sylv
    -0.73
     motorists
    -0.72
    olding
    -0.71
    haar
    -0.71
     shoppers
    -0.69
     popul
    -0.67
     curtains
    -0.67
    ophen
    -0.66
    ampunk
    -0.64
     debunk
    -0.64
    POSITIVE LOGITS
     Played
    1.30
     Became
    1.18
     Career
    1.08
    Joined
    1.06
     Started
    1.03
     Joined
    0.92
     Finished
    0.90
     Former
    0.89
     Injury
    0.88
     coached
    0.86
    Activations Density 0.167%

    No Known Activations