INDEX

Explanations

references to traditional concepts or practices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

ropolis

-0.17

lying

-0.15

bras

-0.15

 Tradition

-0.15

 tradition

-0.15

 Reputation

-0.15

liness

-0.15

hoa

-0.14

aging

-0.14

laus

-0.14

POSITIVE LOGITS

ists

0.43

ist

0.38

ism

0.29

istic

0.28

ista

0.25

izing

0.25

ISTS

0.25

ise

0.25

ized

0.24

isti

0.24

Activations Density 0.031%

references to traditional concepts or practices

No Comments

No Known Activations