INDEX

Explanations

the word "despite" and variations of it, which denote concession or contradiction

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

bps

-0.15

ëĳĺ

-0.14

tras

-0.14

]={↵

-0.14

ssue

-0.14

ÌĢ

-0.13

abouts

-0.13

ãĥ©ãĥ¼

-0.13

 GOODS

-0.13

 Unexpected

-0.13

POSITIVE LOGITS

 being

0.23

 knowing

0.23

 pleas

0.23

 warnings

0.22

 having

0.21

 occasional

0.21

 appearances

0.21

 everything

0.20

 initial

0.20

 claims

0.20

Activations Density 0.032%

the word "despite" and variations of it, which denote concession or contradiction

No Comments

No Known Activations