Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 25633
    Prev
    Next
    INDEX
    Explanations

    forms of the verb "to be" and related variations

    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
    373
    +0.17
    0.8%
    159
    +0.12
    0.6%
    694
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.17
    -0.01
    404
    +0.12
    0.02
    466
    +0.11
    0.02
    Negative Logits
    ãĤ¼ãĤ¦ãĤ¹
    -0.72
    Spread
    -0.70
     needle
    -0.62
    ortium
    -0.62
    SPONSORED
    -0.58
    rical
    -0.54
     Reson
    -0.54
    lehem
    -0.53
    aris
    -0.53
    \<
    -0.52
    POSITIVE LOGITS
    kefeller
    0.90
    issance
    0.80
    hani
    0.74
    kj
    0.73
    hess
    0.72
    earchers
    0.70
    backer
    0.70
    zl
    0.68
    heed
    0.65
    fters
    0.64
    Activations Density 0.031%

    No Known Activations