© 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. 4-RES-SM
    11. 9499
    Prev
    Next
    INDEX
    Explanations

    categories related to various topics or subjects

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    170
    +0.15
    0.9%
    307
    +0.14
    0.8%
    478
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    440
    +0.15
    0.01
    48
    +0.14
    0.01
    197
    +0.12
    0.01
    Negative Logits
    hold
    -1.93
    ments
    -1.60
    bars
    -1.54
    fulness
    -1.53
     center
    -1.53
    ctor
    -1.52
    que
    -1.41
    .\
    -1.40
    gery
    -1.40
     bars
    -1.40
    POSITIVE LOGITS
    Į
    4.01
    į
    3.51
    Ī
    3.33
    ĵ
    3.31
    Ļ
    3.25
    ·¸
    3.19
    ļ
    3.05
    ĥ½
    3.04
    ĺ
    3.04
    ĭ
    3.03
    Activations Density 0.007%

    No Known Activations