© 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. 2-RES-SM
    11. 12875
    Prev
    Next
    INDEX
    Explanations

    specific programming constructs and keywords

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    271
    +0.23
    1.3%
    320
    +0.20
    1.2%
    153
    +0.17
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    320
    +0.23
    0.13
    153
    +0.20
    0.07
    436
    +0.17
    0.07
    Negative Logits
     yours
    -1.66
     mergers
    -1.59
     ours
    -1.52
     divisions
    -1.46
    [^
    -1.43
     solutions
    -1.42
     agrees
    -1.41
     acquisitions
    -1.41
     respectively
    -1.40
     counterparts
    -1.34
    POSITIVE LOGITS
    "}](#
    1.64
    infty
    1.55
    ั
    1.53
    onom
    1.38
    á½IJ
    1.38
    utter
    1.36
    gado
    1.34
    ripp
    1.33
     coli
    1.32
    Ń
    1.31
    Activations Density 3.313%

    No Known Activations