INDEX

Explanations

references to structured content or headings in the document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

oss

-0.17

ones

-0.15

kit

-0.15

aul

-0.15

 Middle

-0.15

istor

-0.15

 worn

-0.15

place

-0.14

ior

-0.14

pic

-0.14

POSITIVE LOGITS

âĨĴâĨĴ

0.20

leo

0.16

erah

0.16

FINE

0.15

 âĨĲ

0.15

 Older

0.15

Ð°Ð»ÑĮÐ½Ðµ

0.15

ï¸

0.15

âĻł

0.14

ASE

0.14

Activations Density 0.005%

references to structured content or headings in the document

No Comments

No Known Activations