INDEX

Explanations

words related to events or actions sequentially or temporally ordered

oai_token-act-pair · gpt-3.5-turbo

instances of the word "after."

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

ahime

-0.84

olds

-0.80

cci

-0.79

arnaev

-0.79

ãĥ¯

-0.78

ouble

-0.77

 guiActiveUn

-0.76

çīĪ

-0.75

JV

-0.74

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.71

POSITIVE LOGITS

 noticing

1.30

 spotting

1.28

 discovering

1.23

 witnessing

1.22

 receiving

1.20

 completing

1.16

 seeing

1.15

 arriving

1.15

 hearing

1.13

 realizing

1.12

Activations Density 0.108%

words related to events or actions sequentially or temporally ordered

instances of the word "after."

No Comments

No Known Activations