© 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. 12885
    Prev
    Next
    INDEX
    Explanations

    HTML document structure and markup elements

    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
    111
    +0.46
    3.2%
    156
    +0.45
    3.1%
    271
    +0.17
    1.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.46
    0.14
    156
    +0.45
    0.10
    71
    +0.17
    0.10
    Negative Logits
    <|outofrange|>
    -2.10
    ↵↵
    -2.10
    ↵
    -2.10
    č↵
    -2.10
    -2.10
    ↵
    -2.10
    č↵
    -2.10
    ↵  ³³³
    -2.10
    -2.10
     ↵
    -2.10
    POSITIVE LOGITS
    DOCTYPE
    1.57
    lig
    1.51
    ael
    1.46
    textstyle
    1.33
    ldots
    1.32
    halt
    1.32
    cdots
    1.31
    ligt
    1.29
    typeof
    1.28
    tfrac
    1.28
    Activations Density 0.052%

    No Known Activations