INDEX

Explanations

phrases indicating the concept of "opposites" or contrasting ideas

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

lac

-0.16

 adipiscing

-0.16

IDD

-0.16

lings

-0.15

lin

-0.15

self

-0.14

istics

-0.14

ling

-0.14

lis

-0.14

iri

-0.14

POSITIVE LOGITS

-sex

0.20

/op

0.19

 extremes

0.18

 extreme

0.17

veau

0.17

 nhau

0.17

 effect

0.17

 direction

0.16

.Toolkit

0.16

-direction

0.15

Activations Density 0.021%

phrases indicating the concept of "opposites" or contrasting ideas

No Comments

No Known Activations