© 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. 37218
    Prev
    Next
    INDEX
    Explanations

    instances of the substring "il"

    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
    373
    +0.13
    0.6%
    238
    +0.13
    0.6%
    106
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    106
    +0.13
    0.02
    238
    +0.13
    0.01
    501
    +0.10
    0.01
    Negative Logits
    ĸļ
    -0.80
    -+-+
    -0.69
    ©¶æ¥µ
    -0.69
    ĪĴ
    -0.65
    olicy
    -0.65
    Ħ¢
    -0.64
    nect
    -0.63
     Ruler
    -0.63
     Citation
    -0.60
     Winds
    -0.57
    POSITIVE LOGITS
    aments
    0.95
     Nadu
    0.91
    ibr
    0.90
    uvian
    0.88
    azar
    0.87
    iberal
    0.86
    icious
    0.84
    ogical
    0.84
    oti
    0.81
    uxe
    0.80
    Activations Density 0.010%

    No Known Activations