© 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. 30525
    Prev
    Next
    INDEX
    Explanations

    references to addiction and its impact on individuals and their families

    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
    326
    +0.23
    1.3%
    235
    +0.13
    0.7%
    444
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.23
    0.11
    451
    +0.13
    0.04
    235
    +0.13
    0.10
    Negative Logits
    soever
    -1.69
     Biosystems
    -1.67
     thereafter
    -1.57
     accordingly
    -1.57
     GmbH
    -1.50
    ));
    -1.48
     Studios
    -1.46
     Lincoln
    -1.42
     regularly
    -1.40
     remotely
    -1.40
    POSITIVE LOGITS
    ace
    1.77
     wrong
    1.62
    ãģªãģĦ
    1.57
    problem
    1.53
    acea
    1.50
    bone
    1.49
     offence
    1.49
    ius
    1.47
     done
    1.46
     problem
    1.45
    Activations Density 7.132%

    No Known Activations