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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & 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. 12155
    Prev
    Next
    INDEX
    Explanations

    terms related to legal restrictions and prohibitions

    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
    75
    +0.12
    0.5%
    197
    +0.10
    0.5%
    153
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    75
    +0.12
    0.03
    421
    +0.10
    0.02
    403
    +0.09
    0.02
    Negative Logits
    bard
    -0.77
    Lenin
    -0.70
    Kings
    -0.69
    tone
    -0.69
    osterone
    -0.68
    ser
    -0.66
    Bee
    -0.66
    arse
    -0.65
    etics
    -0.65
    å°Ĩ
    -0.65
    POSITIVE LOGITS
     anyone
    0.99
     interfering
    0.92
     tampering
    0.91
     anybody
    0.88
     discrimination
    0.88
     any
    0.86
     accessing
    0.84
    ¿½
    0.83
     undue
    0.82
     disclosing
    0.80
    Activations Density 0.037%

    No Known Activations