INDEX

Explanations

phrases related to punctuations and formatting

oai_token-act-pair · gpt-3.5-turbo

punctuation marks, particularly periods and commas

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 occupations

-0.75

 attrition

-0.74

 affili

-0.72

 temperament

-0.69

 conting

-0.69

 elbows

-0.68

 sciences

-0.67

 crossings

-0.65

 settlements

-0.65

 grop

-0.65

POSITIVE LOGITS

 Basically

1.45

Its

1.42

 Essentially

1.40

 Called

1.35

It

1.35

 Featuring

1.33

 Located

1.31

 This

1.29

 Whilst

1.28

 Designed

1.27

Activations Density 0.438%

phrases related to punctuations and formatting

punctuation marks, particularly periods and commas

No Comments

No Known Activations