INDEX

Explanations

the words "End" at the end of phrases

oai_token-act-pair · gpt-3.5-turbo

phrases or sentiments associated with endings or conclusions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

BILITY

-0.63

KER

-0.62

KING

-0.60

hips

-0.59

 prick

-0.59

 thinner

-0.58

king

-0.57

 inherit

-0.57

 spot

-0.56

æ©Ł

-0.56

POSITIVE LOGITS

angered

1.44

owment

1.23

angering

1.21

orse

1.18

ogenous

1.17

ocrine

1.16

urance

1.15

ocrin

1.15

orph

1.11

orses

1.07

Activations Density 0.028%

the words "End" at the end of phrases

phrases or sentiments associated with endings or conclusions

No Comments

No Known Activations