INDEX

Explanations

references to sweetness or pleasantness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

.scalablytyped

-0.20

ivan

-0.17

YLE

-0.16

naments

-0.15

sing

-0.15

ãģĹãģı

-0.14

Ã®t

-0.14

ffset

-0.14

Ð¾ÐºÐ¾Ð½

-0.14

åĸ¶

-0.14

POSITIVE LOGITS

uator

0.15

ened

0.15

ly

0.14

ener

0.14

ening

0.14

iggins

0.14

Globals

0.14

tsx

0.14

 Scho

0.14

acker

0.14

Activations Density 0.012%

references to sweetness or pleasantness

No Comments

No Known Activations