INDEX

Explanations

conditional phrases and questions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

ldr

-0.19

atters

-0.18

ALSE

-0.16

nika

-0.15

}elseif

-0.15

gin

-0.14

ged

-0.14

erno

-0.14

assing

-0.14

ourd

-0.13

POSITIVE LOGITS

 merely

0.18

deaux

0.15

izin

0.15

åıªæĺ¯

0.15

SHIP

0.15

 yoksa

0.15

 just

0.14

 ever

0.14

 îł

0.14

Activations Density 0.030%

conditional phrases and questions

No Comments

No Known Activations