© 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. 13948
    Prev
    Next
    INDEX
    Explanations

    text related to legal or licensing terms

    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
    478
    +0.14
    0.8%
    154
    +0.12
    0.7%
    366
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    450
    +0.14
    0.10
    154
    +0.12
    0.14
    192
    +0.11
    0.04
    Negative Logits
    anza
    -1.68
    uit
    -1.42
    rous
    -1.40
    entional
    -1.39
     together
    -1.38
     season
    -1.37
    racted
    -1.37
    isset
    -1.36
    rique
    -1.35
    ään
    -1.34
    POSITIVE LOGITS
    ±
    1.50
    ¡
    1.49
    ¼
    1.49
    ¥
    1.48
    ĭ
    1.44
    ĸ´
    1.44
    ·
    1.44
     grant
    1.42
    ¸
    1.35
    ī
    1.34
    Activations Density 3.788%

    No Known Activations