© 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. 4-RES-SM
    11. 13233
    Prev
    Next
    INDEX
    Explanations

    terms related to legal and regulatory processes

    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
    21
    +0.26
    1.5%
    397
    +0.13
    0.8%
    377
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    216
    +0.26
    0.01
    375
    +0.13
    0.01
    302
    +0.10
    0.01
    Negative Logits
    Īĺ
    -3.19
    «
    -2.89
    -2.65
    -2.65
    -2.65
    -2.65
    -2.65
    <|outofrange|>
    -2.65
    -2.65
    <|padding|>
    -2.65
    POSITIVE LOGITS
     relates
    1.55
     herein
    1.54
     This
    1.44
     pertains
    1.43
    {
    1.42
    iary
    1.39
     hereinafter
    1.38
     comprises
    1.37
    *](#
    1.36
     These
    1.35
    Activations Density 0.022%

    No Known Activations