INDEX

Explanations

phrases emphasizing important terms or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

him

-0.16

iations

-0.15

xt

-0.15

ientes

-0.14

iet

-0.14

.datab

-0.14

ì²¨ë¶Ģ

-0.13

bum

-0.13

iated

-0.13

imeter

-0.13

POSITIVE LOGITS

İ

0.17

 sheer

0.16

doll

0.15

 both

0.15

of

0.15

0.14

Ð¯Ðº

0.14

both

0.14

JI

0.14

enzhen

0.14

Activations Density 0.010%

phrases emphasizing important terms or concepts

No Comments

No Known Activations