INDEX

Explanations

references to early time periods

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

odyn

-0.16

iyi

-0.15

siz

-0.15

pii

-0.15

ìķł

-0.14

pdev

-0.14

reich

-0.14

atum

-0.14

ignite

-0.14

åłĤ

-0.14

POSITIVE LOGITS

 stages

0.21

-stage

0.17

 stage

0.17

-middle

0.17

stage

0.17

/basic

0.16

gee

0.15

 early

0.15

ling

0.15

Activations Density 0.080%

references to early time periods

No Comments

No Known Activations