INDEX

Explanations

comparisons where one thing is deemed to be superior or preferable to another

oai_token-act-pair · gpt-3.5-turbo

comparative phrases expressing something being better or worse than something else

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

encers

-0.86

arta

-0.84

ãĤ¼ãĤ¦ãĤ¹

-0.78

rones

-0.74

united

-0.74

bard

-0.72

iencies

-0.71

assadors

-0.71

anta

-0.71

pter

-0.70

POSITIVE LOGITS

 having

1.01

 pretending

1.00

 relying

0.99

 anything

0.99

 putting

0.98

 messing

0.97

 slapping

0.96

 guessing

0.96

 letting

0.96

 deleting

0.93

Activations Density 0.095%

comparisons where one thing is deemed to be superior or preferable to another

comparative phrases expressing something being better or worse than something else

No Comments

No Known Activations