INDEX

Explanations

instances of combined or merged concepts

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

ÑĢÐ¾Ð²Ð¸Ñĩ

-0.15

contri

-0.15

-INF

-0.14

 billions

-0.14

_usec

-0.14

 Independent

-0.14

pe

-0.14

 Universe

-0.13

 Mayo

-0.13

 dissoci

-0.13

POSITIVE LOGITS

 single

0.30

single

0.27

/single

0.26

-single

0.25

 SINGLE

0.24

(single

0.21

ToOne

0.21

 Single

0.21

singular

0.20

Single

0.20

Activations Density 0.190%

instances of combined or merged concepts

No Comments

No Known Activations