INDEX

Explanations

phrases related to web form inputs and actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

hab

-0.18

/grpc

-0.16

aeper

-0.14

cade

-0.14

 DalÅ¡ÃŃ

-0.14

 ìłĪ

-0.14

afc

-0.14

aus

-0.14

lew

-0.14

à¥Ģà¤§

-0.13

POSITIVE LOGITS

 %}↵

0.36

 %↵

0.19

%>

0.19

*}

0.18

uku

0.18

%,

0.17

%#

0.16

%-

0.15

raman

0.15

Activations Density 0.005%

phrases related to web form inputs and actions

No Comments

No Known Activations