© 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. 1-RES-SM
    11. 24679
    Prev
    Next
    INDEX
    Explanations

    conclusive statements or recommendations made in the 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/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
    +0.18
    1.0%
    376
    +0.11
    0.6%
    25
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    189
    +0.18
    0.02
    83
    +0.11
    0.02
    309
    +0.11
    0.03
    Negative Logits
    ibles
    -1.78
    ”?
    -1.57
    icles
    -1.52
    oss
    -1.51
    Rev
    -1.51
    "?
    -1.48
    hai
    -1.44
    quel
    -1.41
    rig
    -1.39
    ”—
    -1.38
    POSITIVE LOGITS
    thon
    1.73
    forth
    1.54
     finding
    1.54
     wisdom
    1.53
     obtaining
    1.51
    cing
    1.48
    getting
    1.43
    gni
    1.40
     progress
    1.40
     tricky
    1.39
    Activations Density 0.106%

    No Known Activations