INDEX

Explanations

instances of the word "with."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

itzer

-0.17

IELD

-0.16

RIPT

-0.16

ekim

-0.16

 æĥħ

-0.16

xit

-0.15

cko

-0.14

lope

-0.14

atus

-0.13

OURS

-0.13

POSITIVE LOGITS

oft

0.17

oric

0.16

ess

0.16

inde

0.15

 entirety

0.14

 indeed

0.14

bra

0.14

iesen

0.14

edList

0.14

wol

0.14

Activations Density 0.000%

instances of the word "with."

No Comments

No Known Activations