INDEX

Explanations

titles or mentions of honors, such as "Honorable" or "Honors."

oai_token-act-pair · gpt-3.5-turbo

references to honors or titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

phrine

-0.84

 IMAGES

-0.74

âĹ¼

-0.68

 Hebdo

-0.66

 Panther

-0.65

 Barbarian

-0.65

 Mutant

-0.64

PsyNetMessage

-0.64

Adds

-0.64

Alz

-0.63

POSITIVE LOGITS

orable

1.20

eous

1.09

olulu

1.08

esty

0.94

velength

0.93

ors

0.90

oring

0.89

Hon

0.89

ours

0.88

Hon

0.87

Activations Density 0.005%

titles or mentions of honors, such as "Honorable" or "Honors."

references to honors or titles

No Comments

No Known Activations