© 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. 2-RES-JB
    11. 12760
    Prev
    Next
    INDEX
    Explanations

    words related to things that are doubtful, questionable, or of suspect nature

    oai_token-act-pair · gpt-3.5-turbo

    terms associated with doubt and untrustworthiness

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    olon
    -0.94
    ften
    -0.91
    hner
    -0.86
    ynthesis
    -0.84
    eding
    -0.79
    ummer
    -0.78
    eded
    -0.76
    brate
    -0.76
    OVA
    -0.75
    hower
    -0.75
    POSITIVE LOGITS
     questionable
    1.03
     dubious
    0.95
     legality
    0.86
     suspic
    0.76
     foul
    0.69
     indisc
    0.69
     commodities
    0.68
     amounts
    0.68
     astronomical
    0.68
     unethical
    0.67
    Activations Density 0.012%

    No Known Activations