INDEX

Explanations

adverbs emphasizing a particular quality or characteristic

oai_token-act-pair · gpt-3.5-turbo

words expressing strong negative or positive evaluations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ramid

-0.78

olan

-0.75

anwhile

-0.75

arta

-0.70

owing

-0.70

oan

-0.65

chens

-0.65

otor

-0.65

 Broadcasting

-0.64

 Rapp

-0.64

POSITIVE LOGITS

EngineDebug

0.94

 downright

0.87

 Jagu

0.83

ãĤ¦ãĤ¹

0.78

 uncomfortable

0.74

rontal

0.74

 apprehens

0.72

 awkward

0.72

BuyableInstoreAndOnline

0.71

 embarrassing

0.69

Activations Density 0.006%

adverbs emphasizing a particular quality or characteristic

words expressing strong negative or positive evaluations

No Comments

No Known Activations