INDEX

Explanations

words related to spaces or places referred to as "dens."

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "denial" or "deniers."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ramid

-0.64

 Annotations

-0.63

LES

-0.63

 overdue

-0.62

LER

-0.61

 Mandatory

-0.60

CLR

-0.60

cffffcc

-0.60

 permitting

-0.60

 supper

-0.59

POSITIVE LOGITS

unci

1.48

unciation

1.39

izens

1.32

igration

1.19

ounces

1.15

igrated

1.11

uded

1.10

uclear

1.07

atural

1.05

ational

1.03

Activations Density 0.017%

words related to spaces or places referred to as "dens."

references to the concept of "denial" or "deniers."

No Comments

No Known Activations