© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 22558
    Prev
    Next
    INDEX
    Explanations

    phrases related to repetition of actions or the rediscovery of information

    oai_token-act-pair · gpt-3.5-turbo

    instances of the words "previously" and "already."

    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
    anwhile
    -0.70
     Cohn
    -0.68
     Venice
    -0.66
     Empress
    -0.65
     Philip
    -0.65
     Monitor
    -0.64
     Coulter
    -0.64
     Courier
    -0.64
     Parade
    -0.64
     Rudd
    -0.63
    POSITIVE LOGITS
    existing
    1.35
    existent
    1.20
     existing
    1.08
     unsus
    0.94
     occurring
    0.91
     exist
    0.90
     unseen
    0.88
     discounted
    0.87
    mentioned
    0.86
    established
    0.86
    Activations Density 0.138%

    No Known Activations