INDEX

Explanations

references to interrogation and torture practices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 11-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L11R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.11.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

ydro

-0.16

ä¸Ī

-0.16

 Stevens

-0.15

ais

-0.15

Ð¾Ð±Ð¾Ð²

-0.14

ty

-0.14

.define

-0.14

hij

-0.14

 gent

-0.14

ahlen

-0.14

POSITIVE LOGITS

æľ¯

0.15

 Intelligence

0.15

Kun

0.15

essler

0.15

 Lang

0.14

 Leather

0.14

 Jade

0.14

Lang

0.14

ownik

0.14

 LANG

0.14

Activations Density 0.006%

references to interrogation and torture practices

No Comments

No Known Activations