INDEX

Explanations

references to food, particularly ice cream and dessert items

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

 frying

-0.17

 cooking

-0.16

 ÅŁar

-0.16

 bread

-0.16

oise

-0.16

 Marketable

-0.15

 cooked

-0.15

 underwater

-0.15

 Vine

-0.15

bread

-0.15

POSITIVE LOGITS

ice

0.63

Ice

0.54

ice

0.52

Ice

0.48

ICE

0.45

åĨ°

0.41

ICE

0.40

 scoop

0.39

icy

0.35

sco

0.35

Activations Density 0.040%

references to food, particularly ice cream and dessert items

No Comments

No Known Activations