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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 1757
    Prev
    Next
    INDEX
    Explanations

    timestamps of specific events or messages in a conversation

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    +0.22
    1.0%
    202
    +0.10
    0.4%
    74
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.22
    -0.07
    430
    +0.10
    0.11
    6
    +0.08
    0.09
    Negative Logits
     someday
    -0.72
     nowadays
    -0.63
    asionally
    -0.61
     discriminate
    -0.61
     everyday
    -0.60
     beware
    -0.59
    loo
    -0.58
    Others
    -0.57
    ennett
    -0.57
    vec
    -0.56
    POSITIVE LOGITS
    Reviewer
    1.01
     GMT
    0.71
     partName
    0.70
    quickShipAvailable
    0.69
     UTC
    0.69
    Operation
    0.64
    isode
    0.64
    schild
    0.64
    PsyNetMessage
    0.60
     revised
    0.60
    Activations Density 2.244%

    No Known Activations