© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 32052
    Prev
    Next
    INDEX
    Explanations

    elements related to business practices and customer service

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.2.hook_resid_post
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    198
    +0.27
    1.7%
    400
    +0.23
    1.4%
    186
    +0.22
    1.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    198
    +0.27
    0.14
    400
    +0.23
    0.09
    186
    +0.22
    0.07
    Negative Logits
     Edited
    -1.61
    ftware
    -1.61
    oxford
    -1.55
    yahoo
    -1.53
    gmail
    -1.49
     unpublished
    -1.44
    fered
    -1.44
     Archives
    -1.42
    woke
    -1.42
    à¯ģ
    -1.39
    POSITIVE LOGITS
    standing
    1.58
    ickets
    1.50
     ahead
    1.50
    ections
    1.50
    enough
    1.49
     safe
    1.45
     won
    1.44
    ĻĤ
    1.44
    aging
    1.44
    fitting
    1.42
    Activations Density 2.212%

    No Known Activations