INDEX

Explanations

locations or places, particularly with the mention of "New York."

oai_token-act-pair · gpt-3.5-turbo

instances of letter characters or specific identifiers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Doodle

-0.61

 vomit

-0.59

Magikarp

-0.59

 Blacks

-0.58

 notch

-0.55

 seas

-0.53

 Compass

-0.53

 Carnage

-0.53

ctors

-0.52

 grun

-0.52

POSITIVE LOGITS

ander

0.76

ida

0.76

iculture

0.76

anto

0.75

uton

0.74

heim

0.70

any

0.70

orie

0.69

iotic

0.69

ilateral

0.68

Activations Density 0.099%

locations or places, particularly with the mention of "New York."

instances of letter characters or specific identifiers

No Comments

No Known Activations