INDEX

Explanations

adjectives and phrases related to being at risk or susceptible to harm

oai_token-act-pair · gpt-3.5-turbo

references to vulnerability in various contexts

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

 Collider

-0.76

zos

-0.75

tein

-0.73

OTO

-0.69

aster

-0.68

inth

-0.68

ALSE

-0.67

flu

-0.65

ool

-0.63

oret

-0.63

POSITIVE LOGITS

 vulnerable

1.06

 vulner

0.90

ulnerable

0.87

 adolesc

0.84

 victims

0.81

 vulnerability

0.77

ansas

0.76

 bystanders

0.73

 vulnerabilities

0.71

 tenants

0.71

Activations Density 0.021%

adjectives and phrases related to being at risk or susceptible to harm

references to vulnerability in various contexts

No Comments

No Known Activations