INDEX

Explanations

negative phrases related to communication and responses

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

deo

-0.16

aÅĻ

-0.15

jerne

-0.15

Hopefully

-0.15

ataires

-0.14

 Hopefully

-0.14

ichel

-0.13

istros

-0.13

Ø³ØªÛĮ

-0.13

ARGE

-0.13

POSITIVE LOGITS

 comment

0.34

 commenting

0.27

comment

0.27

 elaborate

0.25

 specify

0.24

 Comment

0.24

Comment

0.24

 rule

0.24

say

0.23

 specifies

0.23

Activations Density 0.082%

negative phrases related to communication and responses

No Comments

No Known Activations