© 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 12288
    9. 8-RES_FS12288-JB
    11. 9806
    Prev
    Next
    INDEX
    Explanations

    terms related to segregation or distinction of items into different groups or categories

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    372
    +0.14
    0.6%
    658
    +0.10
    0.5%
    84
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    658
    +0.14
    0.03
    372
    +0.10
    0.02
    402
    +0.10
    0.02
    Negative Logits
    Ö¼
    -0.69
    eno
    -0.68
    vous
    -0.63
    amac
    -0.62
     Beans
    -0.62
    enos
    -0.61
    ENE
    -0.61
    older
    -0.61
    claimer
    -0.60
     Phelps
    -0.60
    POSITIVE LOGITS
     separate
    0.87
    ively
    0.84
     batches
    0.82
     detach
    0.81
     entities
    0.77
     entity
    0.76
    iating
    0.76
    iates
    0.76
    iations
    0.75
     administrations
    0.72
    Activations Density 0.073%

    No Known Activations