INDEX

Explanations

terms that emphasize inclusivity or participation among diverse groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

ogne

-0.19

relude

-0.15

oba

-0.15

.native

-0.15

izza

-0.15

ocker

-0.15

heck

-0.14

Ì£

-0.14

uitka

-0.14

smarty

-0.14

POSITIVE LOGITS

who

0.30

who

0.25

 involved

0.21

 whom

0.20

 concerned

0.18

iets

0.18

ones

0.17

 kinds

0.16

 segments

0.16

 ages

0.15

Activations Density 0.056%

terms that emphasize inclusivity or participation among diverse groups

No Comments

No Known Activations