INDEX

Explanations

websites or clickable interactions

oai_token-act-pair · gpt-3.5-turbo

references to "Op" or "Op-ed" sections in articles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

åŃĲ

-0.84

ãĤ¼ãĤ¦ãĤ¹

-0.81

ãĥīãĥ©ãĤ´ãĥ³

-0.80

ILLE

-0.79

EngineDebug

-0.74

ARDS

-0.73

ãĤ¨ãĥ«

-0.71

wagen

-0.69

ãģ®å®

-0.68

ãģ¦

-0.68

POSITIVE LOGITS

aque

1.07

Op

0.99

yright

0.99

Op

0.97

osite

0.94

ober

0.93

osta

0.92

ensibly

0.91

orter

0.91

etr

0.91

Activations Density 0.008%

websites or clickable interactions

references to "Op" or "Op-ed" sections in articles

No Comments

No Known Activations