© 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. 1654
    Prev
    Next
    INDEX
    Explanations

    phrases related to investigation and analysis

    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
    348
    +0.12
    0.5%
    40
    +0.11
    0.5%
    624
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    123
    +0.12
    0.06
    22
    +0.11
    0.05
    192
    +0.10
    0.04
    Negative Logits
    BUS
    -0.69
    lly
    -0.68
    ĺ
    -0.66
     Advertisement
    -0.66
    0000000000000000
    -0.64
    gal
    -0.61
    ential
    -0.61
    Fax
    -0.61
    emis
    -0.60
    INGS
    -0.60
    POSITIVE LOGITS
     question
    1.25
    effect
    0.91
    escap
    0.90
    efficiency
    0.86
     front
    0.86
     favor
    0.84
     totality
    0.82
    duc
    0.82
    below
    0.81
     conjunction
    0.80
    Activations Density 0.215%

    No Known Activations