INDEX

Explanations

the word "for" in various contexts

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

iams

-0.16

isku

-0.16

Ø¨ÙĪØ¨

-0.15

asted

-0.14

ä»¥æĿ¥

-0.14

 Ø³Ùģ

-0.14

iam

-0.14

endi

-0.14

leston

-0.14

urum

-0.13

POSITIVE LOGITS

 ever

0.34

 good

0.31

-ever

0.30

ever

0.27

Ever

0.27

 life

0.26

good

0.26

-good

0.26

 Ever

0.24

GOOD

0.21

Activations Density 0.060%

the word "for" in various contexts

No Comments

No Known Activations