© 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. 13425
    Prev
    Next
    INDEX
    Explanations

    words related to regulations or official rules

    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
    +0.16
    0.9%
    203
    +0.14
    0.8%
    470
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    203
    +0.16
    0.02
    410
    +0.14
    0.02
    386
    +0.13
    0.01
    Negative Logits
    llo
    -1.85
    lla
    -1.73
    lli
    -1.58
    anuts
    -1.56
    ÅĦ
    -1.50
     shortest
    -1.48
    ĸ´
    -1.45
    za
    -1.45
    ware
    -1.45
    ylon
    -1.43
    POSITIVE LOGITS
    istered
    2.53
    ulators
    1.96
    ional
    1.88
    isters
    1.78
    ulator
    1.76
    gae
    1.71
    imental
    1.69
    ulatory
    1.64
    num
    1.60
    ulations
    1.54
    Activations Density 0.106%

    No Known Activations