INDEX

Explanations

words related to writing, such as "writes," "wrote," "writing," and "writings."

oai_token-act-pair · gpt-3.5-turbo

occurrences of the phrase "wrote in" followed by context or details about writings

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

isSpecialOrderable

-0.76

ahime

-0.73

adi

-0.71

HUD

-0.67

ItemThumbnailImage

-0.67

OF

-0.66

Kids

-0.66

htar

-0.65

heim

-0.64

ãĥ¯ãĥ³

-0.63

POSITIVE LOGITS

 Bulletin

0.71

 Gazette

0.70

mail

0.70

pmwiki

0.70

ah

0.66

letters

0.66

 Journal

0.66

 Slate

0.65

issance

0.64

 Paste

0.62

Activations Density 0.127%

words related to writing, such as "writes," "wrote," "writing," and "writings."

occurrences of the phrase "wrote in" followed by context or details about writings

No Comments

No Known Activations