INDEX

Explanations

phrases indicating public opinion or survey results

oai_token-act-pair · gpt-3.5-turbo

statements and claims made by respondents in surveys or polls

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĤ´ãĥ³

-0.88

Redditor

-0.85

ovember

-0.75

ntil

-0.73

 extraord

-0.72

Scroll

-0.72

hoff

-0.71

DragonMagazine

-0.70

izons

-0.68

bart

-0.67

POSITIVE LOGITS

 they

1.06

 themselves

0.85

 respondents

0.81

 their

0.77

 unfavorable

0.77

 homosexuality

0.75

yes

0.74

 owning

0.74

 abortion

0.74

 either

0.74

Activations Density 0.072%

phrases indicating public opinion or survey results

statements and claims made by respondents in surveys or polls

No Comments

No Known Activations