INDEX

Explanations

occurrences of the substring "wr" in words

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

ijo

-0.17

ahoma

-0.16

iji

-0.16

ugin

-0.15

 Blocking

-0.14

panse

-0.14

 bombs

-0.14

lep

-0.14

 imperative

-0.14

POSITIVE LOGITS

inkle

0.31

angler

0.28

appings

0.25

wr

0.25

angling

0.23

ith

0.22

astle

0.21

bane

0.21

apture

0.21

angles

0.20

Activations Density 0.005%

occurrences of the substring "wr" in words

No Comments

No Known Activations