INDEX

Explanations

terms related to separation and isolation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

egin

-0.16

ery

-0.16

ãĥ©ãĤ¯

-0.16

yt

-0.15

-ÑĤÐ¾

-0.15

 Alley

-0.15

_EXTENSIONS

-0.15

compass

-0.14

ruz

-0.14

presso

-0.14

POSITIVE LOGITS

/div

0.27

/dist

0.21

 entity

0.20

 entities

0.20

/se

0.20

 distinct

0.19

ively

0.18

-sex

0.18

 DISTINCT

0.18

 biá»ĩt

0.18

Activations Density 0.031%

terms related to separation and isolation

No Comments

No Known Activations