INDEX

Explanations

links or calls to action in written text

oai_token-act-pair · gpt-3.5-turbo

calls to action related to viewing or engaging with content

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

OTOS

-0.56

 sqor

-0.55

Lew

-0.54

RAM

-0.54

 mont

-0.54

é¾įå¥ĳå£«

-0.54

 Tuls

-0.53

TEXT

-0.52

BILL

-0.51

 tremend

-0.51

POSITIVE LOGITS

 Cancel

0.75

%]

0.64

verning

0.59

 Sample

0.58

oslov

0.57

iquette

0.56

ija

0.53

taboola

0.51

igel

0.51

 unfocusedRange

0.50

Activations Density 0.022%

links or calls to action in written text

calls to action related to viewing or engaging with content

No Comments

No Known Activations