© 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. 14001
    Prev
    Next
    INDEX
    Explanations

    references to frameworks, systems, and methodologies in the context of technology and development

    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
    307
    +0.19
    1.1%
    462
    +0.12
    0.7%
    392
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    462
    +0.19
    0.10
    401
    +0.12
    0.05
    179
    +0.11
    0.09
    Negative Logits
    »¿
    -5.85
    ĻĤ
    -5.71
    ·
    -5.33
    Ĭ
    -5.32
    č↵
    -5.21
     ↵
    -5.21
    -5.21
    ↵
    -5.21
    <|outofrange|>
    -5.21
    -5.21
    POSITIVE LOGITS
     capable
    2.77
     featuring
    2.49
     suitable
    2.38
     incorporating
    2.31
     designed
    2.28
    arel
    2.26
     based
    2.25
     called
    2.21
     tailored
    2.20
     suited
    2.15
    Activations Density 1.358%

    No Known Activations