© 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. 1-RES-SM
    11. 25921
    Prev
    Next
    INDEX
    Explanations

    verbs indicating the state or characteristics of subjects

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    23
    +0.43
    2.7%
    351
    +0.10
    0.6%
    485
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    23
    +0.43
    0.12
    8
    +0.10
    0.08
    32
    +0.10
    0.07
    Negative Logits
    works
    -2.16
    ¥
    -2.15
    ª
    -2.06
    posts
    -2.03
    #:
    -1.92
    ista
    -1.90
    ató
    -1.90
    į
    -1.89
     violates
    -1.88
    endas
    -1.80
    POSITIVE LOGITS
     supposed
    1.94
    poons
    1.86
     pronounced
    1.81
     clustering
    1.77
     expected
    1.75
     widespread
    1.72
     widely
    1.67
     mild
    1.67
     toxin
    1.66
     same
    1.63
    Activations Density 0.656%

    No Known Activations