INDEX

Explanations

specific character sequences and formatting elements, particularly unusual symbols and capitalization patterns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

STD

-0.17

 Dale

-0.16

ohl

-0.15

Hop

-0.15

_FT

-0.15

Hop

-0.15

hop

-0.15

ambre

-0.15

_stdio

-0.14

zee

-0.14

POSITIVE LOGITS

 STAR

0.18

star

0.18

æ´Ĺ

0.17

STAR

0.16

 washed

0.16

ãĤ¹ãĤ¿ãĥ¼

0.16

 Star

0.16

 star

0.15

mont

0.15

OUSE

0.15

Activations Density 0.004%

specific character sequences and formatting elements, particularly unusual symbols and capitalization patterns

No Comments

No Known Activations