INDEX

Explanations

names, actions, and descriptive terms

oai_token-act-pair · gpt-3.5-turbo

names of individuals and titles related to conflict or controversial situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ITNESS

-0.95

 Surveillance

-0.67

vant

-0.63

 Monitoring

-0.62

ItemThumbnailImage

-0.57

ItemImage

-0.57

 pestic

-0.56

Org

-0.55

ifted

-0.55

 Nanto

-0.54

POSITIVE LOGITS

 goodbye

1.17

 onto

1.12

 into

1.11

 away

1.03

 apart

0.99

out

0.95

off

0.93

 overboard

0.93

 aside

0.93

 down

0.88

Activations Density 0.680%

names, actions, and descriptive terms

names of individuals and titles related to conflict or controversial situations

No Comments

No Known Activations