© 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. 9068
    Prev
    Next
    INDEX
    Explanations

    references to the net framework or related technologies

    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
    146
    +0.21
    1.2%
    160
    +0.13
    0.8%
    246
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    146
    +0.21
    0.04
    160
    +0.13
    0.02
    158
    +0.13
    0.03
    Negative Logits
     day
    -2.19
     respect
    -1.66
    ondon
    -1.64
    ating
    -1.56
     month
    -1.52
    ising
    -1.51
    ATCH
    -1.48
    hin
    -1.46
    ollowing
    -1.45
     reminder
    -1.42
    POSITIVE LOGITS
    anyahu
    3.32
    flix
    2.75
    tle
    2.68
    works
    2.62
    beans
    2.58
    ting
    2.54
    scape
    2.42
    izens
    2.34
    ty
    2.32
    bsd
    2.17
    Activations Density 0.174%

    No Known Activations