INDEX

Explanations

references to first-time initiatives and project stages

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

beg

-0.15

elt

-0.14

 WATCH

-0.14

rov

-0.14

hem

-0.14

 bergen

-0.14

æ©

-0.14

 acceler

-0.13

axe

-0.13

 Watch

-0.13

POSITIVE LOGITS

riage

0.18

 phase

0.15

uve

0.14

è®

0.14

cle

0.14

bach

0.14

å¿µ

0.14

zÄħ

0.14

nings

0.14

sehen

0.14

Activations Density 0.073%

references to first-time initiatives and project stages

No Comments

No Known Activations