INDEX

Explanations

numerical values (such as numbers, measurements, and quantities)

oai_token-act-pair · gpt-3.5-turbo

occurrences of asterisks or wildcard symbols

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

assetsadobe

-0.84

 Daesh

-0.62

 runaway

-0.58

 humanities

-0.56

ellar

-0.55

 unification

-0.54

 overwhelmed

-0.53

icio

-0.52

 hoax

-0.52

 rubble

-0.52

POSITIVE LOGITS

3.65

*.

2.77

*,

2.71

**

2.62

*)

2.62

*:

2.49

*/

2.30

.*

2.29

)*

2.25

2.13

Activations Density 0.017%

numerical values (such as numbers, measurements, and quantities)

occurrences of asterisks or wildcard symbols

No Comments

No Known Activations