© 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. 4-RES-SM
    11. 10217
    Prev
    Next
    INDEX
    Explanations

    the letter "D" followed by various numbers

    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
    271
    +0.13
    0.8%
    369
    +0.13
    0.7%
    140
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    369
    +0.13
    0.05
    55
    +0.13
    0.07
    140
    +0.11
    0.05
    Negative Logits
    ments
    -1.57
    sembl
    -1.53
    Ĥ
    -1.53
    sec
    -1.51
    pecific
    -1.49
     example
    -1.49
     forward
    -1.47
    į
    -1.43
    eu
    -1.42
     freely
    -1.41
    POSITIVE LOGITS
    OUBLE
    2.51
    aughter
    2.40
    iscuss
    2.38
    ossier
    2.37
    ifference
    2.28
    aughters
    2.25
    irections
    2.19
    irection
    2.09
    ancing
    2.02
    oubt
    1.98
    Activations Density 0.109%

    No Known Activations