INDEX

Explanations

words related to pre-existing or preparatory conditions

oai_token-act-pair · gpt-3.5-turbo

occurrences of the prefix "pre"

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

loo

-0.81

 Slug

-0.74

 Soup

-0.73

 Heights

-0.72

 Throne

-0.70

 Darling

-0.67

 Hearts

-0.66

 ILCS

-0.66

ï¸

-0.66

 Kingdoms

-0.65

POSITIVE LOGITS

requisites

1.14

fecture

1.07

req

0.98

occup

0.93

hend

0.92

ppers

0.92

zn

0.90

fect

0.89

ceed

0.88

achers

0.88

Activations Density 0.016%

words related to pre-existing or preparatory conditions

occurrences of the prefix "pre"

No Comments

No Known Activations