INDEX

Explanations

adjectives related to criticism or judgment

oai_token-act-pair · gpt-3.5-turbo

terms related to misrepresentation or mistakes in characterization and judgment

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

minster

-0.75

negie

-0.71

galitarian

-0.69

 Skydragon

-0.66

Lago

-0.65

ãģ®å®

-0.64

ulhu

-0.62

instein

-0.62

anian

-0.62

 Enhancement

-0.61

POSITIVE LOGITS

vous

0.85

gotten

0.83

 assumptions

0.78

ãĤ®

0.75

ceived

0.74

 mistakes

0.69

 deceive

0.67

untled

0.67

urities

0.67

 misinterpret

0.67

Activations Density 0.141%

adjectives related to criticism or judgment

terms related to misrepresentation or mistakes in characterization and judgment

No Comments

No Known Activations