INDEX

Explanations

words related to communication or advocacy

oai_token-act-pair · gpt-3.5-turbo

references to vocabulary and language usage

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

 Redemption

-0.69

Ô

-0.68

 Reincarn

-0.68

 Mara

-0.65

 Firefly

-0.65

pend

-0.65

 Capitalism

-0.64

 Dying

-0.64

Suz

-0.63

FactoryReloaded

-0.63

POSITIVE LOGITS

voc

1.17

ifer

1.11

abulary

1.02

ationally

0.94

ourses

0.93

oded

0.93

 veter

0.90

ational

0.90

unci

0.88

oder

0.87

Activations Density 0.009%

words related to communication or advocacy

references to vocabulary and language usage

No Comments

No Known Activations