© 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 1536
    9. 8-RES_FS1536-JB
    11. 88
    Prev
    Next
    INDEX
    Explanations

    the word "give" with a strong activation value

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    213
    +0.12
    0.5%
    508
    +0.10
    0.5%
    63
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    217
    +0.12
    0.07
    213
    +0.10
    0.06
    279
    +0.10
    0.07
    Negative Logits
    ansas
    -0.72
    Sphere
    -0.66
    can
    -0.63
    tenance
    -0.59
    timer
    -0.59
    ategy
    -0.59
    pora
    -0.58
    hetto
    -0.58
    uder
    -0.57
    vell
    -0.57
    POSITIVE LOGITS
     rise
    0.99
     birth
    0.98
    aways
    0.95
     refunds
    0.84
     us
    0.84
     chase
    0.83
    away
    0.81
     generously
    0.79
     precedence
    0.76
     them
    0.75
    Activations Density 1.029%

    No Known Activations