INDEX

Explanations

proper nouns or names related to various subjects

oai_token-act-pair · gpt-3.5-turbo

uppercase 'L' characters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 prohibition

-0.68

 boundaries

-0.67

 Subtle

-0.65

éĹĺ

-0.63

 Prohibition

-0.63

ãĥ¼ãĥĨãĤ£

-0.62

fulness

-0.61

ruary

-0.59

 spirited

-0.59

 shattering

-0.58

POSITIVE LOGITS

yrics

1.29

ONDON

1.23

ateral

1.22

aughter

1.21

ifestyle

1.17

incoln

1.16

OWER

1.15

aptop

1.14

ifest

1.14

UX

1.13

Activations Density 0.047%

proper nouns or names related to various subjects

uppercase 'L' characters

No Comments

No Known Activations