INDEX

Explanations

adjectives related to responsibility and irresponsibility

oai_token-act-pair · gpt-3.5-turbo

terms associated with responsibility and irresponsibility

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

 Heard

-0.80

listed

-0.77

bourne

-0.74

iem

-0.72

hew

-0.71

eda

-0.71

brance

-0.69

rooms

-0.68

fields

-0.67

eting

-0.65

POSITIVE LOGITS

 irresponsible

1.11

 responsibly

0.99

responsible

0.78

wcsstore

0.74

 adolesc

0.73

wcs

0.71

 repro

0.71

 responsibility

0.69

 citiz

0.69

æł

0.69

Activations Density 0.012%

adjectives related to responsibility and irresponsibility

terms associated with responsibility and irresponsibility

No Comments

No Known Activations