INDEX

Explanations

numbers and their formatting

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 11-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L11R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.11.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

ely

-0.16

 literature

-0.15

 being

-0.15

ufe

-0.15

 FAST

-0.14

Wnd

-0.14

ajes

-0.14

 forever

-0.14

ry

-0.13

aturas

-0.13

POSITIVE LOGITS

oco

0.15

conds

0.15

utex

0.15

wash

0.14

lad

0.14

ysa

0.14

adil

0.14

anje

0.13

ÎµÏģÎ¹

0.13

 AXIS

0.13

Activations Density 0.116%

numbers and their formatting

No Comments

No Known Activations