INDEX

Explanations

HTML attributes and class names associated with elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

okable

-0.15

Ãły

-0.14

dig

-0.14

dbl

-0.14

åĪ°çļĦ

-0.14

 Ã¡n

-0.14

arters

-0.13

ollider

-0.13

(HWND

-0.13

ìĿ´ìĬĪ

-0.13

POSITIVE LOGITS

clearfix

0.18

col

0.17

 tear

0.16

 active

0.15

ecies

0.15

ui

0.15

Ä±rak

0.15

ql

0.14

ajs

0.14

col

0.14

Activations Density 0.032%

HTML attributes and class names associated with elements

No Comments

No Known Activations