INDEX

Explanations

geographical locations and directional information

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

ples

-0.16

yms

-0.16

InThe

-0.15

edl

-0.15

 Roads

-0.14

roke

-0.14

 Hire

-0.14

adro

-0.13

oref

-0.13

warts

-0.13

POSITIVE LOGITS

 exit

0.23

exit

0.20

ALT

0.19

 ramp

0.19

 Exit

0.19

vi

0.18

 interchange

0.18

Vi

0.17

 Express

0.17

Activations Density 0.042%

geographical locations and directional information

No Comments

No Known Activations