INDEX

Explanations

Phrases or names containing the specific term "ha"

oai_token-act-pair · gpt-3.5-turbo

the presence of the word "ha" in various contexts

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

 margins

-0.74

 Preferences

-0.63

 Conserv

-0.63

 strings

-0.60

 denomin

-0.59

 watch

-0.58

ins

-0.58

 outs

-0.58

 uncomp

-0.58

 scrimmage

-0.57

POSITIVE LOGITS

ha

4.48

hi

1.87

hao

1.78

hu

1.68

hal

1.68

han

1.66

ho

1.62

Ha

1.61

has

1.60

HA

1.60

Activations Density 0.007%

Phrases or names containing the specific term "ha"

the presence of the word "ha" in various contexts

No Comments

No Known Activations