INDEX

Explanations

words related to giving direct instructions or commands

oai_token-act-pair · gpt-3.5-turbo

terms associated with directness and financial transactions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Cheong

-0.61

 Shepard

-0.53

 Keller

-0.52

KR

-0.52

 Hubbard

-0.51

Interstitial

-0.50

 ï¿½

-0.50

Kro

-0.48

 Berks

-0.48

Pwr

-0.46

POSITIVE LOGITS

tenance

1.07

theless

1.03

terday

0.95

withstanding

0.90

usterity

0.83

ardless

0.80

etheless

0.79

yond

0.78

bidden

0.78

itage

0.77

Activations Density 0.434%

words related to giving direct instructions or commands

terms associated with directness and financial transactions

No Comments

No Known Activations