INDEX

Explanations

professional terms or contexts

oai_token-act-pair · gpt-3.5-turbo

references to professional status or expertise across various fields

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĸļ

-0.94

ãĤ¼ãĤ¦ãĤ¹

-0.81

 Strait

-0.74

owship

-0.72

 Halls

-0.70

gow

-0.68

 Sins

-0.66

 Manip

-0.65

 Flavoring

-0.64

 Cavern

-0.64

POSITIVE LOGITS

digy

1.49

verbs

1.14

 wrestler

1.03

actively

0.99

ctor

0.95

ccess

0.93

cedes

0.91

spective

0.90

blems

0.89

strate

0.88

Activations Density 0.017%

professional terms or contexts

references to professional status or expertise across various fields

No Comments

No Known Activations