INDEX

Explanations

specific numbered or ranked components and attributes

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

reh

-0.15

 private

-0.15

aper

-0.15

abet

-0.15

oloj

-0.14

estruct

-0.14

antidad

-0.14

StatusCode

-0.14

 Private

-0.13

dispatch

-0.13

POSITIVE LOGITS

 amongst

0.21

among

0.21

 among

0.20

erus

0.19

 Among

0.17

erner

0.16

essor

0.16

Among

0.15

swick

0.15

ä¾

0.14

Activations Density 0.086%

specific numbered or ranked components and attributes

No Comments

No Known Activations