© 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. 13929
    Prev
    Next
    INDEX
    Explanations

    technical descriptions and specifications related to electrical components and systems

    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
    156
    +0.26
    1.5%
    397
    +0.14
    0.8%
    254
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.26
    0.42
    156
    +0.14
    0.31
    71
    +0.12
    0.29
    Negative Logits
    ivered
    -1.48
    herin
    -1.44
    áĥĶ
    -1.35
    igion
    -1.35
    OVA
    -1.34
    orship
    -1.34
    à±ģ
    -1.32
    ETHOD
    -1.30
    argv
    -1.30
    acin
    -1.30
    POSITIVE LOGITS
    ĩ
    3.36
    IJ
    3.29
    »¿
    3.24
    ¨
    3.19
    ĸ
    3.03
    ħ
    3.02
    ¢
    3.02
    ©
    2.97
    Ķ
    2.95
    ī
    2.86
    Activations Density 0.707%

    No Known Activations