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. 45971
    Prev
    Next
    INDEX
    Explanations

    punctuation marks, especially periods

    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
    679
    +0.19
    0.9%
    430
    +0.12
    0.5%
    640
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    679
    +0.19
    0.09
    6
    +0.12
    0.09
    440
    +0.10
    0.07
    Negative Logits
     transitions
    -0.74
     unequ
    -0.73
     reversible
    -0.71
     extinct
    -0.71
     assassin
    -0.70
     lifes
    -0.69
    ãĤ©
    -0.68
     mosqu
    -0.68
     talent
    -0.67
     dips
    -0.67
    POSITIVE LOGITS
     Specifically
    1.28
     Instead
    1.21
     Asked
    1.19
     Saying
    1.18
     According
    1.17
     Speaking
    1.13
     Rather
    1.11
     Though
    1.11
     Whereas
    1.09
     However
    1.08
    Activations Density 0.249%

    No Known Activations