INDEX

Explanations

phrases or terms related to positions or levels of objects, particularly pertaining to the top or upper part

oai_token-act-pair · gpt-3.5-turbo

elements related to visual or structural components in a document

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

oult

-0.65

Friend

-0.64

phia

-0.64

ruciating

-0.61

isine

-0.61

tesy

-0.61

cffff

-0.61

 anecd

-0.61

»Ĵ

-0.61

illions

-0.60

POSITIVE LOGITS

endment

0.66

selage

0.65

 loader

0.65

 hurdle

0.61

ourse

0.61

loader

0.61

 segment

0.60

 indicator

0.60

hing

0.60

 contender

0.60

Activations Density 0.468%

phrases or terms related to positions or levels of objects, particularly pertaining to the top or upper part

elements related to visual or structural components in a document

No Comments

No Known Activations