INDEX

Explanations

phrases related to recommendations or instructions involving tools or equipment usage

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

ics

-0.16

gross

-0.14

']!='

-0.14

Fav

-0.14

ourg

-0.14

igr

-0.14

ÅĻet

-0.14

æĦ

-0.14

Î±Î¹Î½

-0.14

 ìķĦìĿ´ëĶĶ

-0.13

POSITIVE LOGITS

inters

0.18

acente

0.15

ukes

0.15

llen

0.15

ucha

0.14

ints

0.14

arl

0.14

ivet

0.14

veis

0.14

llib

0.14

Activations Density 0.002%

phrases related to recommendations or instructions involving tools or equipment usage

No Comments

No Known Activations