© 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. 8-RES-JB
    11. 7905
    Prev
    Next
    INDEX
    Explanations

    locations or spatial relationships within a given context

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to grounds for action or reasoning

    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
     (?,
    -0.80
    awei
    -0.71
    Loading
    -0.67
    next
    -0.67
    mate
    -0.66
    amine
    -0.65
    nant
    -0.65
    CTV
    -0.64
    xit
    -0.63
    DEN
    -0.62
    POSITIVE LOGITS
     intervals
    0.72
     insistence
    0.65
     perhaps
    0.65
     notwithstanding
    0.64
    ardless
    0.64
    arently
    0.63
     Nanto
    0.63
    illions
    0.62
     unsus
    0.61
     interval
    0.61
    Activations Density 0.499%

    No Known Activations