INDEX

Explanations

punctuation marks used for emphasis or clarity

oai_token-act-pair · gpt-3.5-turbo

punctuations and question marks within the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 credential

-0.74

Äģ

-0.68

 unle

-0.68

bleacher

-0.67

âĢĳ

-0.66

Ã¨re

-0.66

 privileged

-0.66

Åį

-0.66

 orche

-0.65

 spatial

-0.65

POSITIVE LOGITS

 Anyway

1.40

:(

1.28

 Also

1.25

EDIT

1.15

 Whats

1.15

Anyway

1.14

 Logged

1.13

 Secondly

1.09

:)

1.08

 Lastly

1.08

Activations Density 0.483%

punctuation marks used for emphasis or clarity

punctuations and question marks within the text

No Comments

No Known Activations