INDEX

Explanations

references to food presentation or serving platters

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

ullan

-0.15

omor

-0.15

icism

-0.15

thinkable

-0.15

rier

-0.15

likle

-0.15

icity

-0.14

 DÃŃky

-0.14

sembl

-0.14

_tables

-0.14

POSITIVE LOGITS

aus

0.35

au

0.34

aux

0.32

lets

0.30

forme

0.26

ful

0.26

let

0.26

ware

0.21

ways

0.21

clo

0.21

Activations Density 0.014%

references to food presentation or serving platters

No Comments

No Known Activations