© 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. 12662
    Prev
    Next
    INDEX
    Explanations

    mentions of U.S. states and their respective courts

    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
    71
    +0.20
    1.1%
    410
    +0.14
    0.8%
    369
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    55
    +0.20
    0.03
    145
    +0.14
    0.03
    175
    +0.12
    0.02
    Negative Logits
    '):
    -1.88
    .',
    -1.70
    ?).
    -1.66
    *),
    -1.64
    iph
    -1.62
     etc
    -1.60
    while
    -1.59
    ()),
    -1.58
    .),
    -1.55
    $),
    -1.54
    POSITIVE LOGITS
    Ļª
    3.76
    ĻĤ
    2.88
    ģ
    2.75
    »¿
    2.54
    º
    2.49
    ¿½
    2.20
    ĸ´
    2.18
    ı
    2.09
    £
    2.06
    ¯
    2.06
    Activations Density 0.080%

    No Known Activations