INDEX

Explanations

references to significant messages or statements being conveyed

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 12-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L12R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.12.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

ç½

-0.16

agua

-0.15

_QUOTES

-0.15

elig

-0.14

Inline

-0.14

 leak

-0.14

airo

-0.14

anga

-0.13

ered

-0.13

è¡Ĺéģĵ

-0.13

POSITIVE LOGITS

odem

0.17

stellung

0.16

 message

0.15

istory

0.15

rika

0.14

irection

0.14

 directions

0.14

assing

0.14

 stol

0.14

æµ´

0.14

Activations Density 0.121%

references to significant messages or statements being conveyed

No Comments

No Known Activations