INDEX

Explanations

non-verbal cues or placeholders in the text

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

ĥĿ

-0.17

rette

-0.16

abra

-0.14

owski

-0.14

ilent

-0.14

spy

-0.14

pile

-0.14

etu

-0.13

adesh

-0.13

ÅĤad

-0.13

POSITIVE LOGITS

ãĤ«ãĥ«

0.15

nbsp

0.15

objs

0.14

nda

0.14

abouts

0.13

Ð¼Ñĥ

0.13

 Halloween

0.13

ynom

0.13

.eclipse

0.13

LV

0.13

Activations Density 0.000%

non-verbal cues or placeholders in the text

No Comments

No Known Activations