INDEX

Explanations

phrases indicating medical tests or health-related assessments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

aid

-0.17

Å¡ov

-0.16

ved

-0.15

anel

-0.15

ood

-0.14

uard

-0.14

 Supern

-0.14

aim

-0.14

anking

-0.14

Ĭ¶æĢģ

-0.13

POSITIVE LOGITS

 Ð²ÑģÑĤÑĢÐµ

0.21

 ÐºÐ°ÑĩÐµ

0.19

 ÑģÐ¾ÑħÑĢÐ°

0.19

 ÑģÑĥÑīÐµ

0.18

 histo

0.18

 ÐłÐµÑģÐ¿

0.18

 Ð½Ð°Ð·Ð½Ð°

0.17

 Ð²Ð½ÐµÑĪ

0.17

Phi

0.17

Geo

0.17

Activations Density 0.285%

phrases indicating medical tests or health-related assessments

No Comments

No Known Activations