© 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 3072
    9. 8-RES_FS3072-JB
    11. 3053
    Prev
    Next
    INDEX
    Explanations

    German words related to law enforcement and related activities

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    36
    +0.14
    0.6%
    343
    +0.13
    0.6%
    372
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.14
    0.13
    343
    +0.13
    0.11
    201
    +0.13
    0.08
    Negative Logits
    ted
    -0.73
    wolves
    -0.72
    lasses
    -0.72
    TY
    -0.71
    ty
    -0.69
    meyer
    -0.69
    lings
    -0.69
    ties
    -0.68
    bread
    -0.67
    sticks
    -0.67
    POSITIVE LOGITS
    uthor
    1.38
    ñ
    1.29
    BILITY
    1.25
    vel
    1.14
    ï
    1.09
    ption
    1.08
    ppa
    1.07
    qua
    1.07
    ña
    1.06
    ÅŁ
    1.05
    Activations Density 0.936%

    No Known Activations