INDEX

Explanations

terms related to cancellation

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

ening

-0.17

ych

-0.16

ed

-0.16

edge

-0.15

semb

-0.15

elow

-0.14

 Platz

-0.14

geh

-0.14

yh

-0.14

emple

-0.14

POSITIVE LOGITS

led

0.34

ation

0.27

LED

0.24

AnimationFrame

0.24

ledon

0.20

ations

0.19

icious

0.19

æİī

0.19

/post

0.18

/res

0.18

Activations Density 0.019%

terms related to cancellation

No Comments

No Known Activations