© 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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 0-RES-JB
    11. 11949
    Prev
    Next
    INDEX
    Explanations

    mentions or references to words ending with "mal"

    oai_token-act-pair · gpt-3.5-turbo

    references to "mal" and its variations, indicating a focus on detrimental aspects or negative contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     summons
    -0.70
     Connie
    -0.69
     packing
    -0.68
     bonds
    -0.66
     Syracuse
    -0.63
     Indy
    -0.63
     grips
    -0.62
     YES
    -0.61
     ponies
    -0.60
     relax
    -0.60
    POSITIVE LOGITS
    mal
    4.45
    Mal
    1.90
    MAL
    1.76
    mis
    1.58
    mol
    1.45
     mal
    1.45
     Mal
    1.43
    mic
    1.25
    mia
    1.23
    mist
    1.23
    Activations Density 0.005%

    No Known Activations