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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 24288
    Prev
    Next
    INDEX
    Explanations

    the word "it" in various contexts, particularly when questioning whether certain actions or beliefs are justified or appropriate

    oai_token-act-pair · gpt-4o-miniTriggered 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
    125
    +0.12
    0.5%
    55
    +0.11
    0.5%
    674
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    125
    +0.12
    0.02
    690
    +0.11
    0.02
    674
    +0.10
    0.02
    Negative Logits
    bies
    -0.80
    papers
    -0.80
    wake
    -0.76
    icion
    -0.74
    former
    -0.72
    idth
    -0.71
    transfer
    -0.69
    Topics
    -0.68
    devices
    -0.66
    dit
    -0.65
    POSITIVE LOGITS
     conceivable
    0.99
     possible
    0.95
     unclear
    0.94
     raining
    0.93
     advisable
    0.92
     ironic
    0.92
     impossible
    0.90
     feasible
    0.90
     worth
    0.88
     worthwhile
    0.79
    Activations Density 0.028%

    No Known Activations