INDEX

Explanations

phrases related to conversations or statements made by characters in a narrative

oai_token-act-pair · gpt-3.5-turbo

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

 practition

-0.88

Ident

-0.87

ãĤ´

-0.83

 exting

-0.83

 pione

-0.82

Virgin

-0.80

inki

-0.78

 entreprene

-0.78

 locom

-0.78

 undermin

-0.76

POSITIVE LOGITS

 Dawson

1.18

 Bryant

1.15

 Carpenter

1.15

 Romo

1.13

 Crawford

1.12

 Pryor

1.11

 Chamberlain

1.11

 McKay

1.10

 Hernandez

1.09

 Cunningham

1.06

Activations Density 0.501%

phrases related to conversations or statements made by characters in a narrative

No Comments

No Known Activations