© 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. 10934
    Prev
    Next
    INDEX
    Explanations

    sections and references to further detail within a text

    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
    307
    +0.29
    1.7%
    478
    +0.16
    0.9%
    274
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    77
    +0.29
    0.09
    139
    +0.16
    0.08
    69
    +0.14
    0.08
    Negative Logits
    ī
    -2.73
    Ģ
    -2.40
    ²
    -2.34
    ĨĴ
    -2.34
    ļ
    -2.33
    »
    -2.33
    č↵č↵
    -2.29
    <|outofrange|>
    -2.29
     ↵
    -2.29
    ↵↵
    -2.29
    POSITIVE LOGITS
     detail
    3.04
     terms
    3.01
     relation
    2.36
     conjunction
    2.32
     accordance
    2.25
     regards
    2.21
    ferential
    2.20
    applicable
    2.07
     order
    2.04
    comple
    2.03
    Activations Density 0.554%

    No Known Activations