INDEX

Explanations

high-frequency function words, indicating a focus on grammatical structure and connections within the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

Ã¡li

-0.16

Ã©r

-0.15

ulong

-0.15

Ø§Ø¯ÛĮ

-0.15

äºĴ

-0.14

pill

-0.14

lÃŃ

-0.14

eti

-0.14

æºª

-0.14

Dur

-0.14

POSITIVE LOGITS

Mage

0.16

adera

0.14

 Tort

0.14

 unexpected

0.14

nob

0.14

bach

0.14

paged

0.14

 tort

0.13

Activations Density 0.005%

high-frequency function words, indicating a focus on grammatical structure and connections within the text

No Comments

No Known Activations