INDEX

Explanations

names or specific terms, especially related to popular culture or individuals

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Hod" along with related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Leone

-0.81

Charge

-0.69

terday

-0.67

IZE

-0.66

ngth

-0.64

 Franch

-0.62

 fries

-0.61

Magazine

-0.61

 recharge

-0.60

 Hebdo

-0.59

POSITIVE LOGITS

yssey

1.23

sworth

1.13

ges

1.05

bard

0.99

lers

0.98

ryn

0.97

rob

0.97

rys

0.97

gins

0.96

wig

0.96

Activations Density 0.011%

names or specific terms, especially related to popular culture or individuals

mentions of the name "Hod" along with related terms

No Comments

No Known Activations