INDEX

Explanations

phrases related to threats or potential negative impacts

oai_token-act-pair · gpt-3.5-turbo

references to threats or damage to rights and institutions

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

ãĤ´ãĥ³

-0.79

omer

-0.73

hots

-0.71

owler

-0.71

oaded

-0.70

placed

-0.69

--+

-0.69

arate

-0.68

atoon

-0.68

tackle

-0.68

POSITIVE LOGITS

 integrity

1.53

 livelihood

1.48

 viability

1.47

 credibility

1.31

 wellbeing

1.25

 lives

1.25

 stability

1.24

 validity

1.23

 reliability

1.20

 effectiveness

1.19

Activations Density 0.211%

phrases related to threats or potential negative impacts

references to threats or damage to rights and institutions

No Comments

No Known Activations