INDEX

Explanations

phrases indicating the extent of something, typically using the phrase "all the way to" or "all the way down to"

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a direction or a path

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

imble

-0.70

issan

-0.65

anton

-0.62

ĪĴ

-0.62

onics

-0.61

liv

-0.56

aum

-0.56

ervation

-0.56

»Ĵ

-0.56

uci

-0.55

POSITIVE LOGITS

 down

0.88

 through

0.83

forward

0.78

 back

0.75

points

0.73

 enthusi

0.72

up

0.72

ÙĴ

0.71

 across

0.71

 round

0.70

Activations Density 0.015%

phrases indicating the extent of something, typically using the phrase "all the way to" or "all the way down to"

phrases that indicate a direction or a path

No Comments

No Known Activations