© 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 EvalsExportsAPI 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. 30531
    Prev
    Next
    INDEX
    Explanations

    actions related to maintaining and improving conditions or states

    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
    263
    +0.23
    1.3%
    323
    +0.15
    0.9%
    476
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    263
    +0.23
    0.15
    358
    +0.15
    0.13
    323
    +0.13
    0.07
    Negative Logits
    arily
    -1.57
    Argument
    -1.55
    ]):
    -1.45
     facie
    -1.39
    arity
    -1.34
    rVert
    -1.32
    η
    -1.32
    ]);
    -1.30
    à±į
    -1.30
    analytic
    -1.30
    POSITIVE LOGITS
    mente
    1.58
    \]](
    1.53
    pez
    1.51
    »¿
    1.43
    office
    1.39
    mouth
    1.33
     articles
    1.28
     ownership
    1.28
     instruments
    1.26
    vate
    1.24
    Activations Density 1.208%

    No Known Activations