INDEX

Explanations

phrases related to workplace issues and violations

oai_token-act-pair · gpt-3.5-turbo

references to authority figures or official positions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

isSpecialOrderable

-0.94

 Surviv

-0.80

 Countries

-0.76

 Governments

-0.76

wikipedia

-0.75

thood

-0.71

olkien

-0.70

 Alive

-0.69

 trillions

-0.68

 Consortium

-0.67

POSITIVE LOGITS

 inappropriately

1.21

 improperly

1.20

 complained

1.14

 reprim

1.14

 inappropriate

1.12

 disorderly

1.10

 rude

1.09

 supervisors

1.08

 disruptive

1.08

 complaining

1.08

Activations Density 0.529%

phrases related to workplace issues and violations

references to authority figures or official positions

No Comments

No Known Activations