INDEX

Explanations

words related to compliance and reporting standards

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

 Coverage

-0.15

 coverage

-0.14

_sidebar

-0.14

_dup

-0.14

DEFINE

-0.13

acam

-0.13

 phá»§

-0.13

ormal

-0.13

 lead

-0.13

 elsewhere

-0.13

POSITIVE LOGITS

 attached

0.20

 corresponding

0.19

 assigned

0.19

 accompanies

0.17

 correspond

0.17

 associated

0.17

 tagged

0.17

 Assigned

0.17

attached

0.16

 Correspond

0.16

Activations Density 0.021%

words related to compliance and reporting standards

No Comments

No Known Activations