INDEX

Explanations

words and phrases related to guidelines and instructions

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

zac

-0.14

unc

-0.14

aldi

-0.14

utz

-0.13

asurer

-0.13

ê±¸

-0.13

 Auditor

-0.13

Ã³c

-0.13

syscall

-0.13

Act

-0.13

POSITIVE LOGITS

abin

0.15

edu

0.15

eral

0.15

erval

0.14

afone

0.14

adem

0.14

ayi

0.14

_suite

0.14

³

0.14

Activations Density 0.457%

words and phrases related to guidelines and instructions

No Comments

No Known Activations