© 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. 30630
    Prev
    Next
    INDEX
    Explanations

    terms related to technical documentation and troubleshooting in software 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/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
    156
    +0.44
    2.7%
    111
    +0.14
    0.8%
    496
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    416
    +0.44
    0.01
    369
    +0.14
    0.00
    27
    +0.12
    0.01
    Negative Logits
    tees
    -1.56
    »¿
    -1.50
    ĥ½
    -1.50
     recipient
    -1.50
    ĭ
    -1.44
    ihad
    -1.43
    pired
    -1.43
    ){#
    -1.40
    omitempty
    -1.34
    ophy
    -1.33
    POSITIVE LOGITS
    terday
    1.68
    bridge
    1.40
    \.
    1.38
     onCreate
    1.31
    click
    1.27
    lining
    1.26
    rangle
    1.23
    ERTY
    1.21
    .$
    1.20
    inger
    1.19
    Activations Density 0.037%

    No Known Activations