INDEX

Explanations

phrases indicating actions or events that have occurred or are in progress

oai_token-act-pair · gpt-3.5-turbo

phrases relating to legal or formal actions and their statuses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

inventoryQuantity

-0.85

quickShipAvailable

-0.66

itte

-0.66

Hug

-0.63

 overe

-0.61

Gaming

-0.60

 mates

-0.58

 gross

-0.56

hig

-0.56

Cart

-0.56

POSITIVE LOGITS

ģ«

0.76

ŃĶ

0.75

itored

0.71

 Borough

0.69

},"

0.69

 throughout

0.68

Ĥª

0.68

abouts

0.67

ĺħ

0.66

 ensued

0.66

Activations Density 0.393%

phrases indicating actions or events that have occurred or are in progress

phrases relating to legal or formal actions and their statuses

No Comments

No Known Activations