© 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. 0-RES-JB
    11. 6182
    Prev
    Next
    INDEX
    Explanations

    forum-related terms or phrases

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "forum" in various contexts

    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
    (*
    -0.69
    ochet
    -0.68
    appings
    -0.67
    irst
    -0.65
    served
    -0.63
    apo
    -0.62
    ablish
    -0.61
    llan
    -0.61
     commanded
    -0.61
    leness
    -0.60
    POSITIVE LOGITS
     forum
    3.63
     forums
    2.77
     Forum
    2.28
    forum
    2.21
     Forums
    2.00
    forums
    1.89
     subreddit
    1.53
     moderators
    1.33
     wiki
    1.29
     moderator
    1.29
    Activations Density 0.020%

    No Known Activations