INDEX

Explanations

words related to technology and devices

oai_token-act-pair · gpt-3.5-turbo

descriptors that indicate capability or functionality

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

otaur

-0.81

aeper

-0.77

erva

-0.72

itect

-0.72

arte

-0.71

udeb

-0.65

krit

-0.65

ebus

-0.64

ourning

-0.63

atcher

-0.62

POSITIVE LOGITS

able

0.81

lihood

0.81

ABLE

0.78

âĶĢâĶĢ

0.76

izable

0.75

ible

0.73

ably

0.73

itable

0.73

vable

0.71

nces

0.71

Activations Density 0.023%

words related to technology and devices

descriptors that indicate capability or functionality

No Comments

No Known Activations