INDEX

Explanations

phrases related to a specific person or name, possibly focusing on political figures or events

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names and titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ctors

-0.67

 FIRE

-0.67

glers

-0.66

 Flavoring

-0.64

inki

-0.63

 blunt

-0.60

vae

-0.59

 slash

-0.58

 flames

-0.58

 magnification

-0.58

POSITIVE LOGITS

ettes

0.72

 Lauder

0.70

enne

0.68

steen

0.68

ette

0.67

idge

0.67

Pwr

0.65

endment

0.65

olla

0.65

ISO

0.64

Activations Density 0.135%

phrases related to a specific person or name, possibly focusing on political figures or events

proper nouns, specifically names and titles

No Comments

No Known Activations