© 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. 12310
    Prev
    Next
    INDEX
    Explanations

    words related to a negative impact or stagnation

    oai_token-act-pair · gpt-3.5-turbo

    forms of the verb "to rate" or related derivations

    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
     riot
    -0.65
     naked
    -0.60
     spirited
    -0.58
     Saga
    -0.56
     Warning
    -0.56
     unprotected
    -0.56
     Doctrine
    -0.56
     fumble
    -0.56
     Wond
    -0.55
     writ
    -0.55
    POSITIVE LOGITS
    ates
    4.31
    ATES
    2.81
    ate
    2.54
    ating
    2.21
    ated
    2.20
    ATE
    2.02
    ately
    1.82
    ators
    1.76
    ations
    1.61
    aters
    1.61
    Activations Density 0.014%

    No Known Activations