INDEX

Explanations

terms related to significance and intensity in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 12-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L12R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.12.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

mÃ®

-0.17

ustr

-0.17

nob

-0.15

idge

-0.14

.masks

-0.14

Ð½Ð¾ÑĪ

-0.14

[assembly

-0.14

Ã¶ÃŁe

-0.14

Ð»Ðµ

-0.14

 ÐºÐ¾ÑģÑĤ

-0.14

POSITIVE LOGITS

utter

0.16

 when

0.16

 lorsque

0.15

uer

0.15

urer

0.15

orge

0.14

urt

0.14

fila

0.14

erra

0.13

NF

0.13

Activations Density 0.226%

terms related to significance and intensity in various contexts

No Comments

No Known Activations