INDEX

Explanations

phrases related to physical bending or leaning actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

adem

-0.17

ãĥ³ãĤº

-0.17

erp

-0.16

lette

-0.16

cept

-0.15

ardu

-0.15

roke

-0.15

interop

-0.15

ROKE

-0.15

phe

-0.15

POSITIVE LOGITS

 bend

0.30

 Bend

0.29

 bending

0.28

 bent

0.27

 bends

0.25

 Bent

0.20

-flex

0.19

å¼

0.19

æĽ²

0.18

igo

0.17

Activations Density 0.019%

phrases related to physical bending or leaning actions

No Comments

No Known Activations