INDEX

Explanations

phrases related to fluctuations and changes

oai_token-act-pair · gpt-3.5-turbo

instances of the phrases "ups" and "downs" to signify variations in experiences or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Portuguese

-0.78

Britain

-0.75

Footnote

-0.74

TY

-0.72

Johnson

-0.71

BOOK

-0.70

Missing

-0.69

 Mecca

-0.68

EVA

-0.66

Oak

-0.65

POSITIVE LOGITS

icum

1.06

sembly

1.04

creen

1.01

uits

0.99

aic

0.99

icer

0.97

ides

0.96

ided

0.94

undown

0.93

ourced

0.91

Activations Density 0.011%

phrases related to fluctuations and changes

instances of the phrases "ups" and "downs" to signify variations in experiences or situations

No Comments

No Known Activations