© 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. 13650
    Prev
    Next
    INDEX
    Explanations

    punctuation marks indicating the end of statements or instructions in code

    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
    9
    +0.18
    1.1%
    466
    +0.15
    1.0%
    25
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    9
    +0.18
    0.10
    78
    +0.15
    0.01
    502
    +0.15
    0.08
    Negative Logits
     incentives
    -1.51
     discretion
    -1.40
     contra
    -1.39
     liability
    -1.39
     extent
    -1.36
     origin
    -1.35
    Copyright
    -1.34
     ineffective
    -1.29
     strengths
    -1.27
     involvement
    -1.27
    POSITIVE LOGITS
     ↵
    5.53
    ↵
    5.53
    <|outofrange|>
    5.53
    5.53
    ↵
    5.53
    ↵
    5.53
    5.53
    <|outofrange|>
    5.53
     ↵
    5.53
    ↵↵
    5.53
    Activations Density 0.399%

    No Known Activations