INDEX

Explanations

verbs ending in 'ed' or 'ing' that denote some form of action or behavior

oai_token-act-pair · gpt-3.5-turbo

words associated with reactions or responses, particularly those related to illegal or punitive actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ãĥ¼ãĥ«

-0.67

 weeds

-0.65

 Gmail

-0.64

 EntityItem

-0.60

 dried

-0.58

 ramps

-0.58

 flies

-0.58

ãĥª

-0.58

 paste

-0.58

 consolidated

-0.56

POSITIVE LOGITS

LY

0.79

 against

0.74

soever

0.74

ially

0.73

ean

0.72

fully

0.71

leep

0.71

iate

0.70

aciously

0.70

ignty

0.70

Activations Density 0.172%

verbs ending in 'ed' or 'ing' that denote some form of action or behavior

words associated with reactions or responses, particularly those related to illegal or punitive actions

No Comments

No Known Activations