INDEX

Explanations

various forms of possessive pronouns and references to characters' physical interactions with one another

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

 å¾Ĺ

-0.16

éĴŁ

-0.14

 Bone

-0.14

 backgrounds

-0.14

 Heads

-0.14

logan

-0.14

illi

-0.14

 yourselves

-0.14

erot

-0.14

 Feeling

-0.14

POSITIVE LOGITS

 left

0.25

 right

0.23

 face

0.21

sing

0.20

 arms

0.19

arm

0.19

 index

0.19

 bare

0.18

 jacket

0.18

 hand

0.18

Activations Density 0.133%

various forms of possessive pronouns and references to characters' physical interactions with one another

No Comments

No Known Activations