INDEX

Explanations

phrases related to a specific name or concept, specifically "idi"

oai_token-act-pair · gpt-3.5-turbo

references to idiocy or foolishness in various contexts

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

 guiActiveUnfocused

-0.80

rooms

-0.73

talk

-0.72

DOC

-0.70

deck

-0.66

sticks

-0.65

 rall

-0.65

north

-0.63

 retri

-0.63

POSITIVE LOGITS

idis

0.95

Äĩ

0.94

elson

0.88

ennes

0.86

opsis

0.85

cipl

0.84

idi

0.81

agonal

0.81

emy

0.79

ivari

0.78

Activations Density 0.012%

phrases related to a specific name or concept, specifically "idi"

references to idiocy or foolishness in various contexts

No Comments

No Known Activations