INDEX

Explanations

phrases related to declaring beliefs or stances

oai_token-act-pair · gpt-3.5-turbo

terms related to declarations or statements of belief and profession

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

displayText

-0.70

 inch

-0.64

 captcha

-0.63

 cooker

-0.63

 punch

-0.63

 Sahara

-0.62

 sled

-0.61

 gripping

-0.61

razil

-0.61

 crest

-0.59

POSITIVE LOGITS

orial

1.24

orship

1.17

edly

1.07

edIn

1.06

es

0.93

 profess

0.92

ational

0.90

onte

0.87

ial

0.85

ed

0.85

Activations Density 0.025%

phrases related to declaring beliefs or stances

terms related to declarations or statements of belief and profession

No Comments

No Known Activations