INDEX

Explanations

instances of capitalized acronyms related to technology and organizations

oai_token-act-pair · gpt-3.5-turbo

references to the term "IN" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

onomic

-0.73

ÅĤ

-0.71

uffs

-0.70

ills

-0.69

flies

-0.69

 Oath

-0.69

ãĤ¶

-0.69

unks

-0.69

ãĤª

-0.69

 Diaz

-0.68

POSITIVE LOGITS

TERN

1.28

ER

1.24

CEPT

1.21

LY

1.20

PUT

1.19

OVA

1.18

MENT

1.15

NESS

1.15

LAND

1.12

STON

1.09

Activations Density 0.051%

instances of capitalized acronyms related to technology and organizations

references to the term "IN" in various contexts

No Comments

No Known Activations