© 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. 9315
    Prev
    Next
    INDEX
    Explanations

    adjectives emphasizing negative behaviors or events

    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
    484
    +0.13
    0.6%
    616
    +0.12
    0.5%
    649
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    484
    +0.13
    0.05
    616
    +0.12
    0.05
    436
    +0.11
    0.03
    Negative Logits
    endas
    -0.96
    redients
    -0.77
    appings
    -0.77
    oute
    -0.73
    combe
    -0.73
    ributes
    -0.72
    rea
    -0.72
    lights
    -0.70
    wings
    -0.68
    ickers
    -0.67
    POSITIVE LOGITS
    THING
    1.25
     conceivable
    1.17
     imaginable
    1.08
     given
    1.01
     semblance
    0.92
    WHERE
    0.89
     kind
    0.89
     sane
    0.88
    ONE
    0.86
    where
    0.85
    Activations Density 0.060%

    No Known Activations