INDEX

Explanations

the word "Milan."

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "Milan."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

REE

-0.90

rees

-0.84

URR

-0.73

BOOK

-0.72

cknowled

-0.71

ramid

-0.69

perse

-0.68

ordable

-0.68

yll

-0.66

IELD

-0.66

POSITIVE LOGITS

 Milan

1.08

ovic

1.01

 Marino

0.90

ovich

0.90

ienne

0.87

olini

0.85

ovi

0.82

ese

0.82

opolis

0.82

 Antonio

0.79

Activations Density 0.003%

the word "Milan."

mentions of the word "Milan."

No Comments

No Known Activations