INDEX

Explanations

words associated with identity and distinction, particularly those that highlight cultural or ethnic backgrounds

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tras

-0.16

 __("

-0.15

HAL

-0.14

mate

-0.14

ples

-0.14

enie

-0.14

 Berg

-0.14

_$_

-0.13

Ð½Ð°ÑĢ

-0.13

POSITIVE LOGITS

ellaneous

0.20

patrick

0.18

consin

0.17

.undo

0.17

arro

0.17

lere

0.16

WARDED

0.16

antro

0.16

ulumi

0.15

abella

0.15

Activations Density 0.041%

words associated with identity and distinction, particularly those that highlight cultural or ethnic backgrounds

No Comments

No Known Activations