INDEX

Explanations

Roman numeral references

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

sut

-0.16

isseur

-0.15

aper

-0.15

renc

-0.14

fabs

-0.14

.swagger

-0.14

_dtype

-0.14

Ãªnh

-0.14

ApplicationBuilder

-0.14

amerate

-0.14

POSITIVE LOGITS

III

0.29

ii

0.29

II

0.28

ix

0.27

iii

0.26

ili

0.24

IX

0.23

ll

0.23

vi

0.22

Activations Density 0.012%

Roman numeral references

No Comments

No Known Activations