INDEX

Explanations

vocabulary words related to communication or expression

oai_token-act-pair · gpt-3.5-turbo

words related to vocabulary or verbal expression

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 harmless

-0.68

Ô

-0.66

 Rand

-0.66

 Louie

-0.65

 rule

-0.64

IRD

-0.63

walking

-0.62

Taj

-0.62

iasis

-0.60

 Firefly

-0.60

POSITIVE LOGITS

voc

1.11

abulary

1.09

itive

0.94

ationally

0.93

Voc

0.92

ifer

0.91

ally

0.89

voc

0.89

unci

0.85

ourses

0.85

Activations Density 0.012%

vocabulary words related to communication or expression

words related to vocabulary or verbal expression

No Comments

No Known Activations