INDEX

Explanations

adjectives or adverbs related to seriousness or gravity

oai_token-act-pair · gpt-3.5-turbo

expressions of seriousness and earnestness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

RN

-0.80

AW

-0.72

Trend

-0.71

NUM

-0.71

OTOS

-0.70

 GEAR

-0.66

AMY

-0.65

 Stun

-0.64

 Journals

-0.64

WOR

-0.62

POSITIVE LOGITS

ly

1.06

ness

1.05

lly

0.88

mble

0.87

 earnest

0.85

itial

0.85

lication

0.84

ishly

0.79

ulously

0.77

 endeav

0.76

Activations Density 0.011%

adjectives or adverbs related to seriousness or gravity

expressions of seriousness and earnestness

No Comments

No Known Activations