INDEX

Explanations

mentions of the "dot" term often associated with internet domains or related products

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 20-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L20R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.20.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

ä»ķ

-0.18

ãĥ¼ãĥĳãĥ¼

-0.16

orie

-0.16

hurst

-0.16

oor

-0.15

ladu

-0.15

hold

-0.15

heit

-0.15

orgen

-0.14

esco

-0.14

POSITIVE LOGITS

ting

0.27

fusc

0.19

tingham

0.19

tering

0.18

.Dot

0.18

ter

0.18

(.)

0.18

busters

0.18

.dot

0.18

-dot

0.16

Activations Density 0.017%

mentions of the "dot" term often associated with internet domains or related products

No Comments

No Known Activations