© 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. 1457
    Prev
    Next
    INDEX
    Explanations

    academic paper structure elements like section headings, figure captions, and periods at the end of paragraphs.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @trainjumperx

    academic or research-related content with a focus on brain functions, memory, study designs, and information processing

    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
    373
    +0.13
    0.6%
    423
    +0.10
    0.5%
    587
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.13
    0.08
    174
    +0.10
    0.04
    130
    +0.10
    0.03
    Negative Logits
     Armageddon
    -0.72
     wont
    -0.70
    iHUD
    -0.68
    soDeliveryDate
    -0.68
     creeps
    -0.64
     lett
    -0.64
     Tempest
    -0.64
    terday
    -0.63
    ById
    -0.63
     rul
    -0.62
    POSITIVE LOGITS
    Methods
    1.06
     Methods
    1.04
    Abstract
    1.01
     hypothesized
    0.99
    ivariate
    0.99
    Discussion
    0.99
    RESULTS
    0.98
    ó
    0.97
    FIG
    0.97
     Comparison
    0.97
    Activations Density 0.460%

    No Known Activations