© 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. 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. 8167
    Prev
    Next
    INDEX
    Explanations

    words or phrases related to fear, danger, or intense negative impact

    oai_token-act-pair · gpt-3.5-turbo

    words and phrases related to chilling and eerie effects

    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
    •layer 0 - fear

    No Comments

    Negative Logits
     Obj
    -0.74
     Orient
    -0.73
     mainline
    -0.72
    Object
    -0.70
    Mus
    -0.68
    UA
    -0.66
     faults
    -0.63
    athon
    -0.61
     Uni
    -0.60
     Quarter
    -0.60
    POSITIVE LOGITS
     chilling
    4.00
     disturbing
    1.48
     eerie
    1.44
     chilled
    1.42
     unsettling
    1.38
     terrifying
    1.36
     haunting
    1.36
     harrowing
    1.32
     horrifying
    1.29
     frightening
    1.24
    Activations Density 0.026%

    No Known Activations