© 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. 1-RES-SM
    11. 24351
    Prev
    Next
    INDEX
    Explanations

    negative language associated with misleading or deceptive statements and behaviors

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    +0.18
    1.0%
    47
    +0.16
    0.9%
    504
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    504
    +0.18
    0.12
    56
    +0.16
    -0.06
    320
    +0.14
    0.10
    Negative Logits
    ]>
    -1.50
     newsletters
    -1.48
    :`
    -1.46
    [{
    -1.45
     libraries
    -1.44
     accordingly
    -1.44
    ]];
    -1.44
     aboard
    -1.41
     catalogue
    -1.38
     shelves
    -1.36
    POSITIVE LOGITS
     wrong
    1.97
     innocent
    1.80
    ¼
    1.80
     heck
    1.76
     confrontation
    1.74
     blame
    1.71
     attack
    1.69
     mistake
    1.68
     mis
    1.64
     crap
    1.63
    Activations Density 5.607%

    No Known Activations