© 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. 11559
    Prev
    Next
    INDEX
    Explanations

    words indicating conflict, resistance, or protection against threats

    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
    307
    +0.26
    1.5%
    306
    +0.20
    1.1%
    47
    +0.15
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    306
    +0.26
    0.05
    186
    +0.20
    0.07
    307
    +0.15
    0.08
    Negative Logits
     worth
    -1.86
    worth
    -1.61
    tikzpicture
    -1.58
    aver
    -1.46
    arly
    -1.41
    himself
    -1.41
     ago
    -1.37
    appendix
    -1.36
    igraphy
    -1.35
    otted
    -1.32
    POSITIVE LOGITS
    ĨĴ
    2.50
     shocks
    2.38
    2.28
    ↵
    2.28
    ↵
    2.28
    <|padding|>
    2.28
    ↵↵
    2.28
     ↵
    2.28
    ↵
    2.28
    2.28
    Activations Density 0.666%

    No Known Activations