INDEX

Explanations

terms related to risk reduction and safety mechanisms

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ÐĿÑĥ

-0.13

 Ð¿Ð¾Ð²Ð¸Ð½

-0.13

ansa

-0.13

çĸ

-0.13

 guts

-0.13

agnostic

-0.13

vem

-0.13

linger

-0.13

agra

-0.12

POSITIVE LOGITS

 drafts

0.20

 premature

0.20

sag

0.18

 reflections

0.17

 foreign

0.17

 Ð¿ÑĢÐµÐ¶

0.17

Build

0.17

 ghost

0.16

 undue

0.16

 excessive

0.16

Activations Density 0.214%

terms related to risk reduction and safety mechanisms

No Comments

No Known Activations