© 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 24576
    9. 8-RES_FS24576-JB
    11. 19344
    Prev
    Next
    INDEX
    Explanations

    words related to specific terms or phrases such as "ull" and "Cull" in the text

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    289
    +0.11
    0.5%
    556
    +0.11
    0.5%
    699
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.11
    0.01
    232
    +0.11
    0.01
    289
    +0.10
    0.01
    Negative Logits
    merce
    -0.67
    BOOK
    -0.65
    ãĢIJ
    -0.64
    changes
    -0.63
    PRESS
    -0.62
    Nic
    -0.57
    zig
    -0.57
     Kenya
    -0.56
     Tu
    -0.56
    CE
    -0.56
    POSITIVE LOGITS
    sburg
    1.37
    iard
    1.18
    aby
    1.13
    inx
    1.09
    abal
    1.07
    uminati
    1.05
    igan
    1.05
    s
    1.04
    enic
    1.03
    igans
    1.03
    Activations Density 0.029%

    No Known Activations