INDEX

Explanations

statistical or numerical information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

.getAs

-0.15

AFX

-0.14

å·®

-0.14

ardi

-0.14

 ((__

-0.14

 newPath

-0.14

aktu

-0.14

alsy

-0.14

ÙĲÙĩ

-0.13

ampaign

-0.13

POSITIVE LOGITS

 Nose

0.23

 Appearance

0.21

 Medium

0.21

 deep

0.20

 medium

0.20

 Notes

0.20

 nose

0.19

 notes

0.19

Appearance

0.19

 Deep

0.19

Activations Density 0.017%

statistical or numerical information

No Comments

No Known Activations