INDEX

Explanations

words related to visual appearance

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

ê¸ī

-0.15

ivid

-0.15

endale

-0.14

ê¹Į

-0.14

rawn

-0.14

omor

-0.14

á»©ng

-0.13

ãĥ§

-0.13

¼

-0.13

æŀĿ

-0.13

POSITIVE LOGITS

 nothing

0.22

 suspicious

0.21

 NOTHING

0.21

 exactly

0.21

 alot

0.21

lot

0.20

 identical

0.20

 like

0.19

unc

0.19

nothing

0.19

Activations Density 0.043%

words related to visual appearance

No Comments

No Known Activations