INDEX

Explanations

variations of the letters "w" and certain sentence structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

hang

-0.18

 Levine

-0.18

 puck

-0.17

empt

-0.16

leted

-0.15

les

-0.15

lob

-0.15

AINED

-0.15

legg

-0.15

lets

-0.15

POSITIVE LOGITS

hat

0.21

irtschaft

0.20

rote

0.20

issenschaft

0.20

anj

0.19

istar

0.19

ishes

0.19

tf

0.18

hy

0.18

ipro

0.17

Activations Density 0.142%

variations of the letters "w" and certain sentence structures

No Comments

No Known Activations