© 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 49152
    9. 8-RES_FS49152-JB
    11. 26557
    Prev
    Next
    INDEX
    Explanations

    phrases indicating the concept of "having nothing remaining."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    539
    +0.09
    0.4%
    735
    +0.09
    0.4%
    650
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    650
    +0.09
    0.02
    268
    +0.09
    0.01
    542
    +0.09
    0.01
    Negative Logits
     Cosponsors
    -0.78
    wcsstore
    -0.73
     guid
    -0.70
     Mbps
    -0.66
    antz
    -0.63
     Pengu
    -0.61
     conduc
    -0.61
    ALLY
    -0.61
    cffff
    -0.60
     Facts
    -0.59
    POSITIVE LOGITS
    overs
    1.12
     undone
    0.95
     unanswered
    0.94
     unexpl
    0.88
    onia
    0.79
    eem
    0.77
    oir
    0.77
     untreated
    0.76
     untouched
    0.75
     unresolved
    0.72
    Activations Density 0.015%

    No Known Activations