INDEX

Explanations

phrases related to comparison and contrast

oai_token-act-pair · gpt-3.5-turbo

words or symbols that indicate emphasis or exclamation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

`.

-1.02

`,

-0.95

 alot

-0.92

 doesnt

-0.89

-0.88

-0.84

 dont

-0.83

 didnt

-0.82

..

-0.82

 colourful

-0.74

POSITIVE LOGITS

—

3.38

—"

3.03

)—

2.88

"—

2.81

.—

2.37

—-

2.37

,—

2.29

——

2.02

--

1.94

–

1.94

Activations Density 0.048%

phrases related to comparison and contrast

words or symbols that indicate emphasis or exclamation

No Comments

No Known Activations