INDEX

Explanations

phrases related to crimes or incidents of violence

oai_token-act-pair · gpt-3.5-turbo

references to murders and killings, particularly involving specific individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

taboola

-0.90

wcsstore

-0.85

âĨĳ

-0.82

dding

-0.80

wcs

-0.77

vec

-0.77

iasm

-0.75

Companies

-0.74

hedon

-0.71

owa

-0.71

POSITIVE LOGITS

 Trayvon

1.18

 innocent

1.14

 innoc

1.08

 Freddie

1.02

 unarmed

1.01

 Teresa

1.00

 Cecil

0.99

 Osama

0.97

 teenager

0.95

 Haram

0.94

Activations Density 0.108%

phrases related to crimes or incidents of violence

references to murders and killings, particularly involving specific individuals

No Comments

No Known Activations