© 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. 11393
    Prev
    Next
    INDEX
    Explanations

    closing parentheses in mathematical expressions

    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
    78
    +0.14
    0.8%
    348
    +0.12
    0.7%
    460
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    297
    +0.14
    0.02
    78
    +0.12
    0.00
    191
    +0.11
    0.02
    Negative Logits
    ense
    -1.60
     venue
    -1.40
     501
    -1.39
    cens
    -1.39
    burg
    -1.36
     locality
    -1.34
    hof
    -1.34
    ufact
    -1.33
    sembles
    -1.33
    portation
    -1.33
    POSITIVE LOGITS
    ]{}.
    1.71
     +
    1.51
    »,
    1.50
    ?'
    1.49
     wrt
    1.45
    *(-
    1.43
    *-
    1.42
    ».
    1.42
    ^*(\
    1.39
    .''
    1.39
    Activations Density 0.056%

    No Known Activations