INDEX

Explanations

phrases or concepts related to advancement or improvement to a higher level

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate progression, levels, or transformations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uled

-0.71

 Lists

-0.68

cknowled

-0.66

ãģį

-0.65

argon

-0.62

 lists

-0.62

otted

-0.61

 acknowled

-0.59

imar

-0.58

inflamm

-0.58

POSITIVE LOGITS

 altogether

0.88

 imaginable

0.73

izons

0.68

leon

0.66

iator

0.65

iphany

0.63

vis

0.62

.<

0.61

 entirely

0.61

wcs

0.61

Activations Density 0.219%

phrases or concepts related to advancement or improvement to a higher level

phrases that indicate progression, levels, or transformations

No Comments

No Known Activations