INDEX

Explanations

habits or behaviors mentioned in the text

oai_token-act-pair · gpt-3.5-turbo

references to the concept of habits

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

SAR

-0.73

cross

-0.72

ammy

-0.71

RIS

-0.66

yz

-0.66

zac

-0.64

mberg

-0.64

ndum

-0.64

pta

-0.63

ramid

-0.63

POSITIVE LOGITS

ually

1.15

uated

1.07

 habits

0.99

 habit

0.99

uate

0.99

uation

0.98

uates

0.80

uating

0.78

ucket

0.75

spring

0.71

Activations Density 0.018%

habits or behaviors mentioned in the text

references to the concept of habits

No Comments

No Known Activations