INDEX

Explanations

phrases related to positive audience reactions, such as applause and standing ovation

oai_token-act-pair · gpt-3.5-turbo

expressions of audience reaction and engagement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

terms

-0.93

Ide

-0.79

inventoryQuantity

-0.77

FY

-0.76

soDeliveryDate

-0.75

agate

-0.71

ãĥĪ

-0.70

-0.68

gey

-0.67

 Gutenberg

-0.67

POSITIVE LOGITS

 onlook

1.52

 spectators

1.45

 passers

1.17

 bystanders

1.15

 crowd

1.09

 audience

1.06

 cheering

1.05

 classmates

1.05

 attendees

1.05

 horrified

1.03

Activations Density 0.564%

phrases related to positive audience reactions, such as applause and standing ovation

expressions of audience reaction and engagement

No Comments

No Known Activations