INDEX

Explanations

phrases related to direction or orientation

oai_token-act-pair · gpt-3.5-turbo

references to the concept of being "upside down" or "down" in various contexts

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

cest

-0.87

aph

-0.71

lain

-0.70

"]=>

-0.68

HRC

-0.68

thy

-0.65

 Sturgeon

-0.65

abled

-0.65

inguished

-0.64

 Cosponsors

-0.63

POSITIVE LOGITS

 perspect

0.78

 horizont

0.77

spin

0.75

Iterator

0.72

LOAD

0.69

©¶æ

0.67

 dimensions

0.67

 swoop

0.67

 Dimensions

0.66

 dimension

0.65

Activations Density 0.077%

phrases related to direction or orientation

references to the concept of being "upside down" or "down" in various contexts

No Comments

No Known Activations