INDEX

Explanations

No Explanations Found

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

edef

-0.16

UU

-0.14

ved

-0.14

ruba

-0.14

lude

-0.14

anza

-0.14

lesen

-0.14

 Signature

-0.14

ols

-0.14

DBG

-0.14

POSITIVE LOGITS

igham

0.17

undry

0.16

ãĥ¼ãĥĳãĥ¼

0.15

inger

0.15

fiber

0.15

 disposal

0.15

.getvalue

0.14

avou

0.14

orate

0.14

linger

0.14

Activations Density 0.190%

No Comments

No Known Activations