INDEX

Explanations

phrases indicating things that are not perceived or understood

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 7-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L7R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.7.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

 Insider

-0.16

 mess

-0.15

ele

-0.15

HORT

-0.15

UED

-0.14

iller

-0.14

.portal

-0.14

PLIER

-0.14

 Antar

-0.14

AZE

-0.14

POSITIVE LOGITS

rection

0.16

vant

0.16

 others

0.15

bff

0.15

à¸Ŀ

0.15

inel

0.15

byter

0.14

licken

0.14

etten

0.14

ç¿

0.14

Activations Density 0.098%

phrases indicating things that are not perceived or understood

No Comments

No Known Activations