INDEX

Explanations

phrases related to tragic events

oai_token-act-pair · gpt-3.5-turbo

references to political or social controversies

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

 ability

-0.73

 neighb

-0.70

 duty

-0.69

 eligible

-0.66

 lifes

-0.66

 hatch

-0.64

 undet

-0.64

dra

-0.64

xual

-0.64

 Aval

-0.64

POSITIVE LOGITS

Advertisement

1.35

Whether

1.18

Much

1.16

Some

1.16

Perhaps

1.16

Besides

1.14

But

1.14

Since

1.13

Nevertheless

1.13

Related

1.12

Activations Density 0.755%

phrases related to tragic events

references to political or social controversies

No Comments

No Known Activations