© 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
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 4-RES-SM
    11. 11791
    Prev
    Next
    INDEX
    Explanations

    words related to stereotypes

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    306
    +0.18
    1.0%
    78
    +0.14
    0.8%
    137
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    385
    +0.18
    0.01
    137
    +0.14
    0.01
    349
    +0.13
    0.01
    Negative Logits
    wers
    -1.74
    hips
    -1.73
    hip
    -1.62
     oneself
    -1.59
     yourself
    -1.54
     behalf
    -1.48
     goodbye
    -1.44
    sterdam
    -1.44
     myself
    -1.43
    IER
    -1.42
    POSITIVE LOGITS
    ħ
    4.89
    ij
    4.79
    Ħ
    4.71
    ½
    4.67
    Ĥ¬
    4.55
    Īĺ
    4.44
    »
    4.43
    ģ
    4.42
    Ļ
    4.38
    ¿
    4.31
    Activations Density 0.068%

    No Known Activations