© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 11997
    Prev
    Next
    INDEX
    Explanations

    phrases that describe characteristics and conditions of objects or processes

    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
    +0.19
    1.1%
    348
    +0.17
    1.0%
    21
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    83
    +0.19
    0.06
    23
    +0.17
    0.05
    289
    +0.15
    0.03
    Negative Logits
    ĻĤ
    -2.79
    ↵
    -2.60
    <|outofrange|>
    -2.60
    ↵↵
    -2.60
    ↵
    -2.60
    -2.60
    ↵
    -2.60
    ↵↵
    -2.60
    -2.60
    č↵
    -2.60
    POSITIVE LOGITS
     manually
    1.93
     disposed
    1.82
     conventionally
    1.78
     formed
    1.77
     forcibly
    1.77
     continuously
    1.76
     preferably
    1.76
     xe
    1.75
     substituted
    1.74
     electrically
    1.74
    Activations Density 0.226%

    No Known Activations