INDEX

Explanations

the beginnings of a list or outline

oai_token-act-pair · gpt-3.5-turbo

the phrase "First" indicating a sequence or introduction of points

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

sav

-0.75

bos

-0.72

vor

-0.71

 Gould

-0.68

mbuds

-0.66

âĹ¼

-0.65

ucl

-0.65

utic

-0.63

vim

-0.63

walking

-0.63

POSITIVE LOGITS

 responders

1.01

 baseman

0.88

 glance

0.83

 blush

0.79

 premise

0.78

 impressions

0.77

awa

0.75

ablishment

0.73

 Listen

0.73

 Place

0.71

Activations Density 0.036%

the beginnings of a list or outline

the phrase "First" indicating a sequence or introduction of points

No Comments

No Known Activations