INDEX

Explanations

the word "You."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

Äįan

-0.16

ribbon

-0.15

ritten

-0.15

viron

-0.15

dden

-0.15

baugh

-0.15

phin

-0.15

IFn

-0.14

jours

-0.14

Registry

-0.14

POSITIVE LOGITS

ato

0.15

Ãªs

0.14

wyn

0.14

NB

0.14

 hawk

0.14

IES

0.14

eken

0.14

gap

0.14

.vars

0.14

engers

0.13

Activations Density 0.000%

the word "You."

No Comments

No Known Activations