© 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
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 2021
    Prev
    Next
    INDEX
    Explanations

    phrases indicating someone or something staying in a particular place or situation

    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
    469
    +0.12
    0.5%
    420
    +0.12
    0.5%
    648
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    469
    +0.12
    0.03
    22
    +0.12
    0.03
    654
    +0.09
    0.03
    Negative Logits
    wcsstore
    -0.76
    estic
    -0.74
    entimes
    -0.74
    »Ĵ
    -0.74
    ĪĴ
    -0.70
    catentry
    -0.70
    DragonMagazine
    -0.67
    ¶æ
    -0.67
    elf
    -0.66
    ELF
    -0.66
    POSITIVE LOGITS
     situ
    0.99
     limbo
    0.88
     orbit
    0.83
     bounds
    0.80
     obscurity
    0.80
    oper
    0.78
     infancy
    0.77
    operative
    0.76
     intact
    0.74
     captivity
    0.73
    Activations Density 0.079%

    No Known Activations