INDEX

Explanations

terms related to formal structures or systems

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 5-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L5R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.5.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

/dat

-0.16

ego

-0.15

yb

-0.15

yz

-0.14

abet

-0.14

UME

-0.14

yk

-0.14

_STACK

-0.14

ocking

-0.14

eping

-0.13

POSITIVE LOGITS

dehyde

0.31

ized

0.19

ities

0.18

izes

0.18

/form

0.17

ehler

0.17

ised

0.16

mente

0.15

ovÃŃ

0.15

ize

0.15

Activations Density 0.009%

terms related to formal structures or systems

No Known Activations