INDEX

Explanations

the word "to" used in a statement of preference or intention

oai_token-act-pair · gpt-3.5-turbo

phrases indicating preferences or choices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

LINE

-0.75

worm

-0.64

 Seah

-0.63

 ILCS

-0.62

 Frie

-0.62

 racked

-0.61

 multif

-0.60

 Copyright

-0.59

itzer

-0.59

EMBER

-0.58

POSITIVE LOGITS

asty

0.96

ilitarian

0.75

ugi

0.74

 recommending

0.74

 outright

0.74

 ç¥ŀ

0.73

 seiz

0.71

cellence

0.71

SpaceEngineers

0.67

 defer

0.65

Activations Density 0.130%

the word "to" used in a statement of preference or intention

phrases indicating preferences or choices

No Comments

No Known Activations