INDEX

Explanations

places and venues, especially those related to entertainment and hospitality

oai_token-act-pair · gpt-3.5-turbo

the presence of conjunctions linking different entities and concepts

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

redo

-0.73

 Intent

-0.67

binary

-0.67

rish

-0.67

factor

-0.67

terness

-0.66

iev

-0.65

obin

-0.65

utan

-0.64

XY

-0.64

POSITIVE LOGITS

 gardens

1.20

 restaurants

1.17

 stadiums

1.16

 parks

1.15

 workplaces

1.15

 cafes

1.14

 hotels

1.12

 offices

1.11

 residences

1.11

 kitchens

1.09

Activations Density 0.386%

places and venues, especially those related to entertainment and hospitality

the presence of conjunctions linking different entities and concepts

No Comments

No Known Activations