INDEX

Explanations

variations of the word "ar," possibly indicating a focus on names or terms associated with characters or specific entities in a context

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L1R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.1.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

wat

-0.15

dde

-0.14

aupt

-0.14

ourn

-0.14

anut

-0.14

canf

-0.14

ellig

-0.14

éĥİ

-0.14

andro

-0.14

dsn

-0.13

POSITIVE LOGITS

byss

0.18

Ð»Ð¾Ñĩ

0.16

viewer

0.15

bench

0.15

oned

0.14

gin

0.14

 Volk

0.14

igham

0.14

quia

0.14

rowse

0.14

Activations Density 0.028%

variations of the word "ar," possibly indicating a focus on names or terms associated with characters or specific entities in a context

No Comments

No Known Activations