INDEX

Explanations

words related to support, favoring, or endorsing something or someone

oai_token-act-pair · gpt-3.5-turbo

expressions of public opinion and support for various social and political issues

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĤ¼ãĤ¦ãĤ¹

-0.81

ëĭ

-0.78

DragonMagazine

-0.77

ptoms

-0.73

istical

-0.71

ppa

-0.69

Ñĭ

-0.67

yssey

-0.66

wow

-0.65

 stumble

-0.65

POSITIVE LOGITS

 legalizing

1.32

 stricter

1.15

 tighter

1.13

 repealing

1.13

 abol

1.12

 decriminal

1.09

 expanding

1.09

 banning

1.09

 arming

1.08

 preserving

1.06

Activations Density 0.144%

words related to support, favoring, or endorsing something or someone

expressions of public opinion and support for various social and political issues

No Comments

No Known Activations