INDEX

Explanations

proper nouns or names of companies/organizations

oai_token-act-pair · gpt-3.5-turbo

introductory phrases and words that indicate emphasis or specifics

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĢĤ

-0.63

?).

-0.62

 ðŁĻĤ

-0.61

 Annotations

-0.58

.*

-0.57

ãĤĭ

-0.57

 ãĢĮ

-0.56

.).

-0.56

↵Âł

-0.56

 ----------------

-0.56

POSITIVE LOGITS

withstanding

1.03

resa

1.02

xiety

0.91

icularly

0.89

%"

0.89

usterity

0.87

0.86

odore

0.86

ctions

0.84

ircraft

0.83

Activations Density 0.256%

proper nouns or names of companies/organizations

introductory phrases and words that indicate emphasis or specifics

No Comments

No Known Activations