INDEX

Explanations

words related to susceptibility or exposure to harm

oai_token-act-pair · gpt-3.5-turbo

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zos

-0.72

 Collider

-0.69

inth

-0.68

fare

-0.67

icle

-0.65

furt

-0.64

 fanc

-0.63

exc

-0.63

kos

-0.63

ashi

-0.62

POSITIVE LOGITS

 vulnerable

1.05

ulnerable

0.85

 vulnerability

0.82

 vulner

0.80

 exploited

0.77

 vulnerabilities

0.77

ulnerability

0.77

 susceptible

0.71

ansas

0.69

 victims

0.68

Activations Density 0.040%

words related to susceptibility or exposure to harm

No Comments

No Known Activations