INDEX

Explanations

numerical values referencing different parts of a series

oai_token-act-pair · gpt-3.5-turbo

references to different parts or sections of a series

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ministic

-0.68

jriwal

-0.60

 Cruise

-0.60

avorite

-0.59

anca

-0.58

aukee

-0.57

ITH

-0.57

kefeller

-0.57

 gifted

-0.57

 Predators

-0.56

POSITIVE LOGITS

ridge

1.24

icles

1.20

icle

1.16

icular

1.14

ridges

1.12

nered

1.09

ners

1.09

isans

1.05

ials

1.03

ition

1.00

Activations Density 0.042%

numerical values referencing different parts of a series

references to different parts or sections of a series

No Comments

No Known Activations