INDEX

Explanations

references to entertainment and cultural content

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

chnitt

-0.15

eland

-0.15

ãģĦãĤĭ

-0.15

isson

-0.14

ÑģÐ¾ÑĤ

-0.14

aret

-0.14

æĹ¥ãģ®

-0.14

uras

-0.14

unteer

-0.14

ould

-0.13

POSITIVE LOGITS

tainment

0.16

ertainment

0.16

argo

0.15

adeon

0.14

ARN

0.14

fuse

0.14

ally

0.13

ToLower

0.13

addOn

0.13

iros

0.13

Activations Density 0.020%

references to entertainment and cultural content

No Comments

No Known Activations