© 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 49152
    9. 8-RES_FS49152-JB
    11. 26591
    Prev
    Next
    INDEX
    Explanations

    instances of the word "grace" in various contexts

    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
    459
    +0.15
    0.7%
    255
    +0.13
    0.6%
    572
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    255
    +0.15
    0.01
    259
    +0.13
    0.01
    459
    +0.11
    0.01
    Negative Logits
    ALK
    -0.66
     Regulatory
    -0.61
    ramer
    -0.60
    achu
    -0.60
    aeda
    -0.59
    ERSON
    -0.59
     Lei
    -0.59
     proprietary
    -0.58
     incrim
    -0.58
    industrial
    -0.58
    POSITIVE LOGITS
    fully
    1.71
    fulness
    1.35
    ful
    1.23
    ibility
    1.05
    FUL
    1.05
    full
    1.01
    ifully
    0.90
    ously
    0.87
    ibly
    0.86
    giving
    0.85
    Activations Density 0.007%

    No Known Activations