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. 26051
    Prev
    Next
    INDEX
    Explanations

    phrases related to social responsibility and support for communities

    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
    87
    +0.15
    0.8%
    756
    +0.10
    0.5%
    762
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.15
    0.06
    570
    +0.10
    0.02
    184
    +0.09
    0.03
    Negative Logits
    isSpecialOrderable
    -0.64
    etheless
    -0.62
     Canaver
    -0.60
    terday
    -0.60
     aback
    -0.59
    soDeliveryDate
    -0.57
    catentry
    -0.55
    Fram
    -0.54
    [/
    -0.53
     outlining
    -0.52
    POSITIVE LOGITS
     themselves
    1.23
    their
    0.95
     their
    0.95
     oneself
    0.84
     THEIR
    0.81
     careers
    0.78
    igrate
    0.74
     subsistence
    0.72
    selves
    0.70
    tein
    0.70
    Activations Density 0.344%

    No Known Activations