INDEX

Explanations

words ending with "am."

oai_token-act-pair · gpt-3.5-turbo

variations of the word "am."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 sober

-0.65

 kinderg

-0.62

 stakes

-0.61

 affairs

-0.60

flush

-0.57

 ______

-0.55

 checkout

-0.54

 Kers

-0.54

 puberty

-0.54

POSITIVE LOGITS

nesty

1.23

bitious

1.21

endment

1.17

sterdam

1.14

ilies

1.11

azing

1.10

munition

1.09

ilial

1.07

pler

1.06

pering

1.03

Activations Density 0.045%

words ending with "am."

variations of the word "am."

No Comments

No Known Activations