INDEX

Explanations

words related to programming codes or syntax

oai_token-act-pair · gpt-3.5-turbo

colons and associated information or lists following them

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alog

-0.61

orth

-0.58

eware

-0.57

leased

-0.57

manent

-0.56

entle

-0.56

',"

-0.55

aceous

-0.54

eks

-0.51

,'"

-0.49

POSITIVE LOGITS

<|endoftext|>

1.26

↵↵

1.26

↵

1.23

 https

1.20

 http

1.13

↵Âł

1.12

https

0.85

pic

0.84

"[

0.84

 namely

0.80

Activations Density 0.154%

words related to programming codes or syntax

colons and associated information or lists following them

No Comments

No Known Activations