INDEX

Explanations

phrases related to negative action or events

oai_token-act-pair · gpt-3.5-turbo

the conjunction "and" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ãĤ´ãĥ³

-0.65

pee

-0.60

pai

-0.58

uper

-0.57

rawdownloadcloneembedreportprint

-0.57

 Latest

-0.56

eaturing

-0.54

 newcom

-0.54

Vert

-0.54

stad

-0.54

POSITIVE LOGITS

 that

1.54

that

1.36

 THAT

1.24

 That

1.07

That

1.06

 therefore

0.98

 urged

0.89

 hence

0.82

 furthermore

0.82

 doubted

0.81

Activations Density 0.326%

phrases related to negative action or events

the conjunction "and" in various contexts

No Comments

No Known Activations