INDEX

Explanations

components related to inspection protocols

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

-0.17

hq

-0.17

pad

-0.17

ording

-0.15

Ã¨me

-0.15

Ø±

-0.15

eras

-0.15

PP

-0.15

press

-0.15

ances

-0.15

POSITIVE LOGITS

ovich

0.22

hton

0.21

oth

0.21

loid

0.21

oy

0.20

ervers

0.19

lying

0.18

oi

0.18

 resent

0.18

unct

0.18

Activations Density 0.296%

components related to inspection protocols

No Comments

No Known Activations