INDEX

Explanations

the adjective "moderate" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to moderation in terms of political perspectives or behaviors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

RIP

-0.77

AMES

-0.75

arium

-0.72

 CODE

-0.71

TYPE

-0.70

 Myster

-0.69

yang

-0.69

 Pione

-0.69

Sport

-0.69

verified

-0.68

POSITIVE LOGITS

 moderate

1.11

erate

1.05

moderate

1.04

yip

0.98

 moderates

0.84

 minded

0.84

 leaning

0.83

 medi

0.79

 independents

0.77

 moder

0.77

Activations Density 0.010%

the adjective "moderate" in various contexts

references to moderation in terms of political perspectives or behaviors

No Comments

No Known Activations