© 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. 30875
    Prev
    Next
    INDEX
    Explanations

    references to religious titles and the papacy

    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
    +0.23
    1.3%
    174
    +0.15
    0.8%
    421
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    59
    +0.23
    0.03
    410
    +0.15
    0.05
    174
    +0.10
    0.05
    Negative Logits
    APTER
    -1.43
    ñor
    -1.37
     minor
    -1.35
    phia
    -1.34
     Appellees
    -1.29
     median
    -1.28
     simple
    -1.27
     principle
    -1.25
     probative
    -1.23
    sofar
    -1.22
    POSITIVE LOGITS
    Ĩ
    2.40
    Ģ
    2.31
    ĩ
    2.11
    Ĵ
    2.10
    ĻĤ
    2.07
    Ĭ
    1.95
    ettes
    1.94
    ©
    1.92
    ¹
    1.90
    Ĥ
    1.85
    Activations Density 0.313%

    No Known Activations