© 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. 31871
    Prev
    Next
    INDEX
    Explanations

    hyperlinks or URLs within 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/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
    478
    +0.17
    1.0%
    451
    +0.14
    0.8%
    159
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    29
    +0.17
    0.01
    451
    +0.14
    0.01
    53
    +0.13
    0.02
    Negative Logits
    manship
    -1.99
    hips
    -1.84
    oxacin
    -1.68
     (\<
    -1.64
    atics
    -1.58
    ders
    -1.56
    hip
    -1.56
     controls
    -1.55
     medalists
    -1.49
    onies
    -1.46
    POSITIVE LOGITS
    °
    2.24
    ¶
    2.14
    https
    2.11
    ij
    2.11
    typeof
    2.10
    Ĥ¬
    2.09
    į
    2.08
    ĵ
    2.07
    ±
    2.03
    ://
    2.02
    Activations Density 0.067%

    No Known Activations