INDEX

Explanations

words related to terms with the abbreviation "AT" or "AT" itself

oai_token-act-pair · gpt-3.5-turbo

references to the acronym "AT."

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

perse

-0.71

ordering

-0.69

ppelin

-0.66

ingen

-0.62

theless

-0.60

doms

-0.59

hold

-0.59

oser

-0.59

 cultivated

-0.59

place

-0.58

POSITIVE LOGITS

ICAN

1.25

RON

1.03

terson

1.01

hetically

1.01

rix

0.99

ELY

0.96

ting

0.96

TERN

0.94

itudes

0.93

rice

0.93

Activations Density 0.016%

words related to terms with the abbreviation "AT" or "AT" itself

references to the acronym "AT."

No Comments

No Known Activations