INDEX

Explanations

references to internet domains and email addresses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 16-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L16R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.16.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

Ã©n

-0.13

oload

-0.13

ropa

-0.13

eon

-0.13

ipment

-0.13

StringEncoding

-0.13

 ëĵľë¦½ëĭĪëĭ¤

-0.13

trÃ©al

-0.12

 Vere

-0.12

roup

-0.12

POSITIVE LOGITS

bidden

0.19

adays

0.16

ÑĢÐ°Ð

0.16

 mÃ¤nner

0.15

ourcem

0.15

Phi

0.14

OrCreate

0.14

aceous

0.14

acco

0.14

AXB

0.14

Activations Density 0.188%

references to internet domains and email addresses

No Comments

No Known Activations