© 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. 19902
    Prev
    Next
    INDEX
    Explanations

    phrases related to research focus and expert opinions

    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
    481
    +0.17
    0.8%
    408
    +0.10
    0.5%
    226
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    384
    +0.17
    0.06
    393
    +0.10
    0.06
    703
    +0.10
    0.04
    Negative Logits
    kees
    -1.03
    ummer
    -0.88
    ivas
    -0.83
    ilk
    -0.81
    urs
    -0.79
    blems
    -0.79
    ppo
    -0.78
    weet
    -0.77
    itled
    -0.77
    Happy
    -0.76
    POSITIVE LOGITS
     supplemented
    1.02
     ingenuity
    0.95
     intuition
    0.91
     sheer
    0.89
     techniques
    0.88
     expertise
    0.87
     persuasion
    0.85
     negligence
    0.84
     circumst
    0.84
     aided
    0.83
    Activations Density 0.336%

    No Known Activations