INDEX

Explanations

the word "where"

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

questions or statements inquiring about a specific location

oai_token-act-pair · gpt-3.5-turbo

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

yi

-0.81

ATURE

-0.78

asts

-0.74

ilus

-0.69

agers

-0.69

ME

-0.66

vous

-0.66

astics

-0.66

emed

-0.64

ems

-0.64

POSITIVE LOGITS

abouts

1.34

upon

1.12

fore

1.03

soever

0.82

 else

0.80

ver

0.71

 exactly

0.69

velt

0.67

with

0.65

Dat

0.65

Activations Density 0.049%

the word "where"

questions or statements inquiring about a specific location

No Comments

No Known Activations