© 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 EvalsExportsAPI 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. 9835
    Prev
    Next
    INDEX
    Explanations

    references to intimate or sexual contexts

    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
    369
    +0.19
    1.1%
    18
    +0.16
    0.9%
    494
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    18
    +0.19
    0.03
    494
    +0.16
    0.02
    172
    +0.14
    0.02
    Negative Logits
     behalf
    -1.70
     pregnant
    -1.66
    sis
    -1.63
    quarters
    -1.63
     course
    -1.62
     authors
    -1.58
     generality
    -1.56
     importance
    -1.55
     stomach
    -1.53
    veolar
    -1.51
    POSITIVE LOGITS
    Ĵ
    2.88
    ²
    2.66
    IJ
    2.59
    ³
    2.54
    ª
    2.44
    ı
    2.34
    ĻĤ
    2.32
    ĺ
    2.32
    ¯
    2.30
    Ļª
    2.27
    Activations Density 0.097%

    No Known Activations