INDEX

Explanations

modifiers for sorting or categorization of data

oai_token-act-pair · gpt-3.5-turbo

phrases indicating categorization and sorting by specific criteria

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

hend

-0.78

iron

-0.77

ãĤ£

-0.71

rats

-0.70

DEF

-0.70

uala

-0.69

aphael

-0.69

kee

-0.69

lings

-0.68

hern

-0.68

POSITIVE LOGITS

 nationality

1.31

 ethnicity

1.28

 gender

1.25

 geography

1.25

 geographic

1.19

 geographical

1.16

 category

1.10

age

1.07

 denomination

1.05

 genders

1.04

Activations Density 0.161%

modifiers for sorting or categorization of data

phrases indicating categorization and sorting by specific criteria

No Comments

No Known Activations