INDEX

Explanations

phrases related to workplace issues or environments

oai_token-act-pair · gpt-3.5-turbo

references to the workplace

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

iasis

-0.72

ados

-0.72

Ã¡s

-0.71

beam

-0.70

quet

-0.68

aciously

-0.68

adoes

-0.67

oline

-0.63

olid

-0.63

roe

-0.63

POSITIVE LOGITS

 ethic

0.78

 Advantage

0.74

urized

0.70

 environment

0.70

 hazard

0.69

 wellness

0.68

restling

0.68

Safe

0.66

halla

0.64

places

0.64

Activations Density 0.014%

phrases related to workplace issues or environments

references to the workplace

No Comments

No Known Activations