INDEX

Explanations

the repeated use of the word "with."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 5-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L5R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.5.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

UIL

-0.15

ÏĨÎ®

-0.14

setVisibility

-0.14

 bÃ¬

-0.14

Ð½ÑĮ

-0.13

cape

-0.13

anh

-0.13

ÑĢÑĥÐ±

-0.13

wechat

-0.13

ofil

-0.13

POSITIVE LOGITS

 Rosen

0.16

 accompanying

0.15

rien

0.15

 being

0.15

yor

0.15

alÄ±ÄŁÄ±

0.14

hausen

0.14

 attendant

0.14

ä¸Ģèµ·

0.14

azard

0.14

Activations Density 0.017%

the repeated use of the word "with."

No Comments

No Known Activations