INDEX

Explanations

the letter 't' appearing as a single token

oai_token-act-pair · gpt-3.5-turbo

instances of the letter 't'

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 rounding

-0.68

Vie

-0.66

 Hitman

-0.66

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.66

 tightening

-0.65

 grooming

-0.64

 nonex

-0.64

 exha

-0.64

 destro

-0.63

ĪĴ

-0.63

POSITIVE LOGITS

ribune

1.42

urtle

1.41

utorial

1.38

weet

1.31

ournament

1.31

otally

1.30

uesday

1.29

alks

1.26

ravis

1.26

itles

1.25

Activations Density 0.029%

the letter 't' appearing as a single token

instances of the letter 't'

No Comments

No Known Activations