INDEX

Explanations

numerical sequences, specifically dates and counts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ruba

-0.16

hower

-0.16

onse

-0.16

aper

-0.15

æĪ´

-0.15

ÄĽÅ¾

-0.14

soles

-0.14

pNet

-0.14

Ã©ri

-0.14

bih

-0.14

POSITIVE LOGITS

 Cong

0.14

Immediate

0.14

Cong

0.14

.office

0.14

-tm

0.14

plx

0.14

 háº¡ng

0.13

(Render

0.13

oho

0.13

|↵

0.13

Activations Density 0.019%

numerical sequences, specifically dates and counts

No Comments

No Known Activations