© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 429
    Prev
    Next
    INDEX
    Explanations

    proper nouns and specific places

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    447
    +0.16
    0.7%
    669
    +0.10
    0.4%
    67
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    121
    +0.16
    0.05
    264
    +0.10
    0.04
    350
    +0.09
    0.04
    Negative Logits
    ahime
    -0.89
    bis
    -0.77
    ramid
    -0.74
    Scale
    -0.68
     conco
    -0.67
     pollut
    -0.66
    format
    -0.64
     decom
    -0.63
     guiActiveUnfocused
    -0.62
     unsolved
    -0.62
    POSITIVE LOGITS
     myself
    1.14
     representatives
    1.08
     clergy
    1.08
     spouses
    1.01
     businessmen
    1.01
     celebrities
    1.00
     retirees
    1.00
     civilians
    1.00
     elderly
    0.99
     journalists
    0.97
    Activations Density 0.575%

    No Known Activations