INDEX

Explanations

names of people or characters, particularly surnames

oai_token-act-pair · gpt-3.5-turbo

names of individuals and organizations related to various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ãĥĥãĥĪ

-0.83

UGE

-0.78

isle

-0.72

icate

-0.71

erest

-0.69

eree

-0.65

 Moor

-0.65

 obesity

-0.63

 blanket

-0.63

using

-0.63

POSITIVE LOGITS

FTWARE

0.96

manship

0.95

enegger

0.87

sonian

0.80

0.76

lington

0.75

BALL

0.75

hoff

0.74

Technical

0.73

bourg

0.70

Activations Density 0.022%

names of people or characters, particularly surnames

names of individuals and organizations related to various contexts

No Comments

No Known Activations