INDEX

Explanations

words associated with historical or cultural references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

 Sentence

-0.15

reur

-0.15

 Exit

-0.15

icode

-0.14

arty

-0.14

ano

-0.14

 bubble

-0.14

adero

-0.13

urile

-0.13

æĹ

-0.13

POSITIVE LOGITS

 root

0.20

 Proto

0.20

 Middle

0.20

 roots

0.19

 Late

0.19

Middle

0.18

 cogn

0.18

 Forms

0.17

Forms

0.17

 stem

0.17

Activations Density 0.041%

words associated with historical or cultural references

No Comments

No Known Activations