INDEX

Explanations

numerical references and structured lists in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

cai

-0.15

opi

-0.15

 jack

-0.14

ìĹ¼

-0.14

egt

-0.14

rodu

-0.14

appable

-0.13

Ø§Ø«

-0.13

ulos

-0.13

QUEST

-0.13

POSITIVE LOGITS

0.19

0.16

soever

0.15

hatt

0.14

.cf

0.14

Ð¾Ð±Ð¾Ð²

0.14

:)

0.13

mtx

0.13

 Hicks

0.13

omatic

0.13

Activations Density 0.026%

numerical references and structured lists in the text

No Comments

No Known Activations