INDEX

Explanations

references to the concept of being "straight" or related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 28-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L28R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.28.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

sein

-0.17

quets

-0.15

ials

-0.15

##_

-0.15

-vous

-0.14

soever

-0.14

ingen

-0.14

him

-0.14

nbsp

-0.14

éĮ²

-0.14

POSITIVE LOGITS

ened

0.34

forward

0.33

ening

0.32

away

0.32

-forward

0.29

ener

0.28

eners

0.28

 forward

0.25

est

0.25

aways

0.25

Activations Density 0.011%

references to the concept of being "straight" or related terms

No Comments

No Known Activations