INDEX

Explanations

instances of numerical values or years

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 11-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L11R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.11.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

indsight

-0.17

ihn

-0.16

ibel

-0.16

ibur

-0.16

ibs

-0.15

 Late

-0.15

elsea

-0.15

asion

-0.15

quot

-0.14

kn

-0.14

POSITIVE LOGITS

ieÅ¼

0.17

ationally

0.16

poses

0.15

ç®Ģ

0.14

REEN

0.14

elves

0.14

 porr

0.13

 tavs

0.13

 Clover

0.13

INDOW

0.13

Activations Density 0.016%

instances of numerical values or years

No Comments

No Known Activations