INDEX

Explanations

numeric codes or identifiers commonly found in bibliographic references or documentation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

ruz

-0.15

 GenÃ§

-0.15

è¼

-0.14

hana

-0.14

/trunk

-0.14

 Rough

-0.14

dojo

-0.13

ãĤ¾

-0.13

ersen

-0.13

Ð¾Ð½Ðµ

-0.13

POSITIVE LOGITS

0.17

eker

0.16

.son

0.16

ÑĪÐºÐ¸

0.14

okable

0.14

IMPORT

0.14

Ã¡no

0.14

ipi

0.14

ichick

0.14

Ãłnh

0.14

Activations Density 0.010%

numeric codes or identifiers commonly found in bibliographic references or documentation

No Comments

No Known Activations