INDEX

Explanations

places or locations within a document

oai_token-act-pair · gpt-3.5-turbo

punctuation and commas in the text

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

stead

-0.68

hess

-0.62

 Almighty

-0.60

TON

-0.59

ton

-0.58

tx

-0.58

Tut

-0.58

Ack

-0.57

egg

-0.57

 immunity

-0.57

POSITIVE LOGITS

imi

0.76

umbnails

0.66

¾

0.66

ngth

0.65

rahim

0.65

imming

0.64

²

0.63

etheless

0.63

Ķ

0.63

ãĤ¦

0.62

Activations Density 0.924%

places or locations within a document

punctuation and commas in the text

No Comments

No Known Activations