© 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. 26708
    Prev
    Next
    INDEX
    Explanations

    references to quality assurance and related operational codes

    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
    12
    +0.19
    1.1%
    271
    +0.17
    1.0%
    354
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    479
    +0.19
    0.06
    13
    +0.17
    0.07
    354
    +0.15
    0.09
    Negative Logits
    ·
    -1.62
    į
    -1.60
    ¸
    -1.52
    «
    -1.50
    »¿
    -1.49
    ische
    -1.47
    ľ
    -1.47
    ¶
    -1.45
    Ĥ
    -1.45
    dere
    -1.38
    POSITIVE LOGITS
    ensively
    1.48
     bundles
    1.40
    nets
    1.34
    chemy
    1.34
    ãĥ¼ãĥ«
    1.33
    iece
    1.32
    ector
    1.29
     instantly
    1.27
    icles
    1.27
    oire
    1.26
    Activations Density 1.713%

    No Known Activations