INDEX

Explanations

phrases related to physical actions, especially involving conflict or confrontation

oai_token-act-pair · gpt-3.5-turbo

instances of the word "to" and its related phrases indicating actions or intentions

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

 Done

-0.81

cedented

-0.78

 Balanced

-0.71

 Printed

-0.69

Bench

-0.64

ultimate

-0.62

today

-0.61

 Needs

-0.61

marked

-0.60

Added

-0.59

POSITIVE LOGITS

 intervene

1.31

 bribe

1.28

 assassinate

1.27

 persuade

1.25

 convince

1.25

 revive

1.25

 negotiate

1.23

 infiltrate

1.19

 flee

1.17

 steal

1.15

Activations Density 0.090%

phrases related to physical actions, especially involving conflict or confrontation

instances of the word "to" and its related phrases indicating actions or intentions

No Comments

No Known Activations