© 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. 9558
    Prev
    Next
    INDEX
    Explanations

    quoted speech or dialogue in the text

    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
    130
    +0.12
    0.7%
    23
    +0.12
    0.7%
    263
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    130
    +0.12
    0.06
    179
    +0.12
    0.04
    89
    +0.10
    0.03
    Negative Logits
    :`
    -2.01
    -1.82
    ↵
    -1.82
    <|outofrange|>
    -1.82
    -1.82
    č↵
    -1.82
    ↵
    -1.82
    <|outofrange|>
    -1.82
    čč
    -1.82
    ↵↵↵
    -1.82
    POSITIVE LOGITS
    UX
    1.59
     ends
    1.57
    iph
    1.46
    ambo
    1.45
     Beg
    1.40
     features
    1.36
    shaped
    1.35
    amba
    1.34
    uously
    1.34
    }}'
    1.30
    Activations Density 0.206%

    No Known Activations