© 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. 1686
    Prev
    Next
    INDEX
    Explanations

    phrases expressing positive sentiment or excitement

    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
    36
    +0.19
    0.9%
    263
    +0.12
    0.5%
    393
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.19
    0.03
    263
    +0.12
    0.02
    284
    +0.10
    0.01
    Negative Logits
     ¶
    -0.74
    é¾įå¥ij士
    -0.70
     ILCS
    -0.67
    76561
    -0.64
     Oswald
    -0.64
    apons
    -0.62
    sylvania
    -0.62
    CI
    -0.60
    \<
    -0.60
    STD
    -0.59
    POSITIVE LOGITS
    ils
    0.99
     coincidence
    0.96
     lot
    0.87
    uras
    0.84
     wonderful
    0.84
     bunch
    0.83
     heck
    0.82
     lovely
    0.82
     hypocr
    0.81
    kward
    0.79
    Activations Density 0.027%

    No Known Activations