INDEX

Explanations

the word "nose."

oai_token-act-pair · gpt-3.5-turbo

the suffix 'ose' in words

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

angible

-0.67

rators

-0.66

DOC

-0.66

enberg

-0.66

 ILCS

-0.64

DERR

-0.63

cdn

-0.62

ertodd

-0.61

ufact

-0.61

Ont

-0.61

POSITIVE LOGITS

velt

1.04

cond

0.92

idon

0.92

eker

0.88

lect

0.85

OPLE

0.81

lihood

0.80

ppel

0.80

ï¸

0.78

ffic

0.76

Activations Density 0.016%

the word "nose."

the suffix 'ose' in words

No Comments

No Known Activations