INDEX

Explanations

references to political and legal figures or processes

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

ixed

-0.14

 variety

-0.13

@author

-0.13

 cover

-0.13

itus

-0.13

ãģ¨æĢĿãģĨ

-0.13

ftar

-0.13

oti

-0.12

ogn

-0.12

Ter

-0.12

POSITIVE LOGITS

 earlier

0.29

 last

0.29

 previously

0.28

 yesterday

0.25

 recently

0.23

 hasn

0.23

Earlier

0.22

 Earlier

0.21

 hadn

0.20

 Previously

0.20

Activations Density 0.231%

references to political and legal figures or processes

No Comments

No Known Activations