INDEX

Explanations

sequences of names and possibly certain programming-related terms in the context of code

oai_token-act-pair · gpt-3.5-turbo

entities, particularly names of people and their roles

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

["

-0.70

cffff

-0.70

-0.69

hom

-0.67

 tremend

-0.65

ãĥīãĥ©

-0.65

ãĤ¼ãĤ¦ãĤ¹

-0.65

edom

-0.64

obyl

-0.63

ricular

-0.61

POSITIVE LOGITS

olkien

0.77

/>

0.68

 Weekly

0.64

leted

0.63

ommel

0.62

 Photo

0.62

ertodd

0.60

 Huss

0.59

 Folder

0.59

>)

0.59

Activations Density 0.104%

sequences of names and possibly certain programming-related terms in the context of code

entities, particularly names of people and their roles

No Comments

No Known Activations