INDEX

Explanations

phrases indicating causing a sensation or controversy

oai_token-act-pair · gpt-3.5-turbo

situations or contexts that elicit strong emotional responses

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

vironments

-0.72

 Mavericks

-0.72

 Granted

-0.67

etheless

-0.66

 Defenders

-0.66

olescent

-0.66

girlfriend

-0.65

edge

-0.64

 Everywhere

-0.64

izons

-0.64

POSITIVE LOGITS

 stir

1.40

 stirred

1.05

 stirring

1.01

 Stir

0.96

fry

0.95

rings

0.90

ONSORED

0.86

 nood

0.83

rer

0.82

rers

0.79

Activations Density 0.008%

phrases indicating causing a sensation or controversy

situations or contexts that elicit strong emotional responses

No Comments

No Known Activations