INDEX

Explanations

phrases indicating common knowledge or widely shared information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

-0.15

uters

-0.15

obot

-0.14

icious

-0.14

ron

-0.14

itably

-0.14

uat

-0.14

eday

-0.14

ullo

-0.14

inh

-0.13

POSITIVE LOGITS

istrovstvÃŃ

0.16

 TCHAR

0.15

zew

0.14

okit

0.14

Ø«ÙĦ

0.14

IsRequired

0.14

dÃ¼l

0.14

à¹ģà¸¥

0.14

ÑĪÐ¸

0.13

Ø§ÙģÛĮ

0.13

Activations Density 0.071%

phrases indicating common knowledge or widely shared information

No Comments

No Known Activations