INDEX

Explanations

expressions that involve direct address or second-person pronouns

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

Pocket

-0.14

é¨

-0.14

 PROCUREMENT

-0.14

à¤µà¤°

-0.14

aption

-0.14

 kuru

-0.14

Ð»Ð°ÑĪ

-0.13

gmt

-0.13

steder

-0.13

 recru

-0.13

POSITIVE LOGITS

Ð°Ð¶

0.17

 searching

0.17

 looking

0.16

 Sell

0.16

 requirements

0.15

 requirement

0.15

0.14

 dream

0.14

uba

0.14

 answered

0.14

Activations Density 0.206%

expressions that involve direct address or second-person pronouns

No Comments

No Known Activations