© 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. 20195
    Prev
    Next
    INDEX
    Explanations

    instances of false accusations or claims

    oai_token-act-pair · gpt-3.5-turbo

    instances of false accusations or claims

    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
     teamwork
    -0.68
    strength
    -0.67
    soDeliveryDate
    -0.65
     Hunger
    -0.64
     practical
    -0.63
     foreseeable
    -0.63
     durability
    -0.62
     versatile
    -0.61
    atable
    -0.61
    isation
    -0.61
    POSITIVE LOGITS
     falsely
    2.89
     wrongly
    2.82
     mistakenly
    2.55
     errone
    2.55
     incorrectly
    2.45
     inappropriately
    2.41
     unfairly
    2.40
     improperly
    2.29
     unlawfully
    2.09
     prematurely
    1.98
    Activations Density 0.036%

    No Known Activations