INDEX

Explanations

words related to injuries or potential harm

oai_token-act-pair · gpt-3.5-turbo

keywords related to injuries, harm, and their consequences

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

last

-0.85

achment

-0.78

soDeliveryDate

-0.76

itta

-0.73

atar

-0.73

ãĤ¨

-0.71

liam

-0.71

ku

-0.70

yssey

-0.67

adle

-0.67

POSITIVE LOGITS

nor

1.36

 anymore

1.17

 whatsoever

1.13

 slightest

0.81

 anybody

0.75

 anything

0.75

 laure

0.74

 anywhere

0.72

Laughs

0.68

 except

0.68

Activations Density 0.398%

words related to injuries or potential harm

keywords related to injuries, harm, and their consequences

No Comments

No Known Activations