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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 12427
    Prev
    Next
    INDEX
    Explanations

    specific job-related keywords and corporate terms, particularly in the context of recruitment and SAP-related content

    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
    156
    +0.15
    0.9%
    362
    +0.15
    0.8%
    369
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    173
    +0.15
    0.04
    369
    +0.15
    0.02
    410
    +0.13
    0.08
    Negative Logits
    asted
    -1.56
     behalf
    -1.48
    rys
    -1.37
    ished
    -1.37
     mistaken
    -1.35
    asting
    -1.33
    raising
    -1.33
     browser
    -1.33
    être
    -1.32
     vom
    -1.32
    POSITIVE LOGITS
    Ł
    1.82
    ftware
    1.82
    chn
    1.77
    ¢
    1.71
    Ħ
    1.69
    PECT
    1.69
    Īĺ
    1.68
    ¯
    1.67
    ª
    1.65
    ĸ
    1.63
    Activations Density 0.833%

    No Known Activations