INDEX

Explanations

phrases indicating strong emotions or reactions

oai_token-act-pair · gpt-3.5-turbo

intensifiers or expressions emphasizing a strong sentiment

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nings

-0.67

 Flavoring

-0.63

 looms

-0.61

amac

-0.60

 coincides

-0.60

ulia

-0.59

 eviction

-0.59

 glances

-0.59

å¥³

-0.58

works

-0.57

POSITIVE LOGITS

bered

1.16

ooo

1.07

oths

1.04

oooo

1.02

oooooooo

0.96

zin

0.96

othes

0.93

othe

0.91

oooooooooooooooo

0.89

 much

0.85

Activations Density 0.063%

phrases indicating strong emotions or reactions

intensifiers or expressions emphasizing a strong sentiment

No Comments

No Known Activations