© 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. 37238
    Prev
    Next
    INDEX
    Explanations

    terms related to structures and organization in various contexts

    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
    87
    +0.34
    1.8%
    373
    +0.13
    0.7%
    481
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.34
    0.08
    87
    +0.13
    -0.01
    575
    +0.11
    0.06
    Negative Logits
    姫
    -0.68
    POSE
    -0.63
    ophers
    -0.59
    ipolar
    -0.54
    tsky
    -0.54
    EGIN
    -0.52
    yss
    -0.50
    å¸
    -0.50
    gar
    -0.50
    osate
    -0.50
    POSITIVE LOGITS
    tyard
    0.54
    lessly
    0.52
    hesion
    0.50
    plates
    0.50
    ariat
    0.50
     (/
    0.49
    ciation
    0.48
    itent
    0.48
     circa
    0.47
    izes
    0.47
    Activations Density 0.384%

    No Known Activations