INDEX

Explanations

phrases indicating maximizing or optimizing actions

oai_token-act-pair · gpt-3.5-turbo

expressions of the ability or potential to do something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

furt

-0.82

BuyableInstoreAndOnline

-0.75

 Lauder

-0.72

 Debor

-0.65

EUR

-0.58

gement

-0.58

 groundbreaking

-0.56

ERY

-0.56

agram

-0.56

ELD

-0.55

POSITIVE LOGITS

 afford

1.25

 muster

1.22

 manage

1.06

 possibly

0.98

 tolerate

0.97

 feas

0.93

 handle

0.93

 conceive

0.93

 imagine

0.89

 cram

0.88

Activations Density 0.045%

phrases indicating maximizing or optimizing actions

expressions of the ability or potential to do something

No Comments

No Known Activations