INDEX

Explanations

references to the word "Du"

oai_token-act-pair · gpt-3.5-turbo

references to the name "Du" 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 @ 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ISC

-0.73

oulos

-0.71

ãģĨ

-0.71

giving

-0.71

ULE

-0.69

 Recent

-0.66

 Organizations

-0.63

 Including

-0.63

aceutical

-0.63

Aph

-0.63

POSITIVE LOGITS

Du

1.02

pee

0.93

ente

0.91

alogy

0.88

pling

0.87

Pont

0.86

alog

0.85

Du

0.85

iste

0.82

plin

0.81

Activations Density 0.004%

references to the word "Du"

references to the name "Du" in various contexts

No Comments

No Known Activations