INDEX

Explanations

references to physical strength or strength-related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

asca

-0.18

otts

-0.17

OGLE

-0.16

quin

-0.15

ikan

-0.15

bage

-0.15

/xhtml

-0.14

clarations

-0.14

ê

-0.14

ponder

-0.14

POSITIVE LOGITS

ens

0.27

ening

0.27

ener

0.25

ened

0.25

holds

0.22

 weakness

0.22

 Weak

0.22

/power

0.21

weak

0.21

ning

0.21

Activations Density 0.024%

references to physical strength or strength-related concepts

No Comments

No Known Activations