INDEX

Explanations

variations of the letter "h" in various contexts

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

leet

-0.16

idl

-0.16

ãĥ¼ãĥł

-0.16

haz

-0.15

Ø§ÙĪØ±ÛĮ

-0.15

vere

-0.15

arest

-0.15

unga

-0.15

 ä¸

-0.14

isty

-0.14

POSITIVE LOGITS

inz

0.21

alt

0.21

inter

0.20

ellig

0.20

och

0.18

aft

0.18

oft

0.17

atte

0.17

asser

0.17

afen

0.17

Activations Density 0.009%

variations of the letter "h" in various contexts

No Comments

No Known Activations