INDEX

Explanations

repeated references to the word "thing."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 22-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L22R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.22.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

thin

-0.17

lek

-0.15

ned

-0.15

licht

-0.15

sites

-0.15

ogne

-0.15

ised

-0.14

aptor

-0.14

ted

-0.14

ates

-0.14

POSITIVE LOGITS

ummy

0.24

/people

0.23

ToDo

0.23

ummies

0.20

 happening

0.20

/person

0.18

0.16

 Happ

0.16

 happen

0.16

Activations Density 0.066%

repeated references to the word "thing."

No Comments

No Known Activations