© 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. 8-RES-JB
    11. 11529
    Prev
    Next
    INDEX
    Explanations

    references to visual perception and interpretation

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "see" and its variations, indicating a focus on perception and observation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    orney
    -0.73
     Dickinson
    -0.66
     mayors
    -0.61
     wheelchair
    -0.61
    hero
    -0.61
     Doyle
    -0.59
    jee
    -0.58
     Drift
    -0.58
    ipt
    -0.58
    travel
    -0.57
    POSITIVE LOGITS
     happening
    0.85
     constitutes
    0.75
    ãĤ´
    0.74
     transpired
    0.72
     unfold
    0.72
     happen
    0.71
     beforehand
    0.69
    ilk
    0.68
     wrought
    0.68
     Learned
    0.67
    Activations Density 0.199%

    No Known Activations