INDEX

Explanations

sequences and structured elements

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

arem

-0.16

ushima

-0.15

amat

-0.14

kal

-0.14

aren

-0.14

Ã©l

-0.14

Ã©f

-0.14

argin

-0.14

ivol

-0.13

POSITIVE LOGITS

_CALLBACK

0.15

 ëĶ°

0.15

Callback

0.14

 callback

0.14

 Commonwealth

0.14

Ã§ek

0.14

 Callback

0.14

ÏĮ

0.14

 Ø¹Øµ

0.14

Bas

0.14

Activations Density 0.024%

sequences and structured elements

No Comments

No Known Activations