INDEX

Explanations

phrases that suggest recommending or providing additional 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

quirer

-0.18

plevel

-0.17

duct

-0.15

sian

-0.15

emark

-0.15

sembles

-0.15

elsey

-0.14

Ĥ

-0.14

ky

-0.14

elsius

-0.14

POSITIVE LOGITS

getting

0.29

cing

0.29

ums

0.27

ced

0.25

 context

0.24

ged

0.24

 those

0.24

bes

0.24

give

0.23

 instance

0.23

Activations Density 0.089%

phrases that suggest recommending or providing additional information

No Comments

No Known Activations