© 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. 305
    Prev
    Next
    INDEX
    Explanations

    random and disconnected phrases with occasional references to social media and politics

    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
    6
    +0.17
    0.8%
    640
    +0.14
    0.6%
    454
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.17
    0.20
    640
    +0.14
    0.14
    679
    +0.10
    0.14
    Negative Logits
     occupation
    -0.84
     migration
    -0.83
     organis
    -0.82
     emanc
    -0.80
     employment
    -0.79
    ciplinary
    -0.79
     objective
    -0.77
     displacement
    -0.76
    ilitation
    -0.76
     transition
    -0.76
    POSITIVE LOGITS
     Anyway
    1.64
     Seriously
    1.56
     ;)
    1.54
    Anyway
    1.50
     Nope
    1.48
     Yeah
    1.44
     Oops
    1.41
     Hmm
    1.40
     Turns
    1.40
     Yep
    1.38
    Activations Density 0.952%

    No Known Activations