INDEX

Explanations

mentions of specific locations and organizations

oai_token-act-pair · gpt-3.5-turbo

the frequent mention of the word "the," often in relation to the United States and associated contexts

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

finals

-0.66

itta

-0.62

izoph

-0.62

opsis

-0.61

ashi

-0.60

ullah

-0.60

reements

-0.59

?????

-0.59

ratulations

-0.59

owners

-0.58

POSITIVE LOGITS

 helm

1.27

 behest

1.26

 forefront

1.19

 moment

1.18

 outset

1.17

 expense

1.11

 disposal

1.02

 discretion

0.92

 time

0.85

 highest

0.84

Activations Density 0.098%

mentions of specific locations and organizations

the frequent mention of the word "the," often in relation to the United States and associated contexts

No Comments

No Known Activations